JP7215243B2 - image coding device - Google Patents

Description

The present invention relates to image encoding and decoding techniques for dividing an image into blocks and performing prediction.

In image encoding and decoding, an image to be processed is divided into blocks each of which is a set of a predetermined number of pixels, and each block is processed. Encoding efficiency is improved by dividing into appropriate blocks and appropriately setting intra-prediction and inter-prediction.

In video encoding/decoding, inter-prediction, which predicts from encoded/decoded pictures, improves encoding efficiency. Patent Literature 1 describes a technique of applying affine transformation in inter prediction. In moving images, it is not uncommon for an object to undergo deformation such as enlargement/reduction and rotation.

JP-A-9-172644

However, since the technique of Patent Document 1 involves image conversion, there is a problem that the processing load is large. SUMMARY OF THE INVENTION In view of the above problems, the present invention provides a low-load and efficient encoding technique.

To solve the above problems, an image coding apparatus according to an aspect of the present invention acquires a plurality of control points having motion vectors, determines reference positions of sub-blocks, and based on the control points and the reference positions, A motion compensation prediction unit that derives a motion vector of the sub-block and performs motion compensation on a sub-block basis, wherein the motion compensation prediction unit divides the sub-block into non-square sub-blocks that are rectangular areas that are not square. , the reference position for deriving the motion vector of the sub-block is the center position of one of the square regions obtained by dividing the non-square sub-block into a plurality of square regions.

According to the present invention, highly efficient image encoding/decoding processing can be realized with a low load.

1 is a block diagram of an image coding device according to an embodiment of the present invention; FIG. 1 is a block diagram of an image decoding device according to an embodiment of the present invention; FIG. FIG. 11 is a flowchart for explaining an operation of dividing a treeblock; FIG. FIG. 4 is a diagram showing how an input image is divided into treeblocks; FIG. 4 is a diagram for explaining z-scan; It is a figure which shows the division|segmentation shape of a block. 4 is a flowchart for explaining an operation of dividing a block into four; 4 is a flowchart for explaining an operation of dividing a block into two or three; This is the syntax for expressing the shape of block division. It is a figure for demonstrating intra prediction. FIG. 4 is a diagram for explaining reference blocks for inter prediction; A syntax for expressing coding block prediction modes. FIG. 4 is a diagram showing the correspondence between syntax elements and modes related to inter-prediction; FIG. 4 is a diagram for explaining affine transformation motion compensation of two control points; FIG. 4 is a diagram for explaining affine transformation motion compensation of three control points; 2 is a block diagram of a detailed configuration of inter prediction section 102 in FIG. 1. FIG. FIG. 17 is a block diagram of a detailed configuration of a normal motion vector predictor mode derivation unit 301 of FIG. 16; FIG. 17 is a block diagram of a detailed configuration of a normal merge mode derivation unit 302 of FIG. 16; 17 is a flowchart for explaining normal predicted motion vector mode derivation processing of a normal predicted motion vector mode deriving unit 301 of FIG. 16; FIG. 10 is a flowchart showing a processing procedure of a normal motion vector predictor mode derivation process; FIG. FIG. 10 is a flowchart for explaining the procedure of normal merge mode derivation processing; FIG. 3 is a block diagram of a detailed configuration of an inter prediction unit 203 in FIG. 2; FIG. FIG. 23 is a block diagram of a detailed configuration of a normal motion vector predictor mode derivation unit 401 of FIG. 22; FIG. 23 is a block diagram of a detailed configuration of a normal merge mode derivation unit 402 of FIG. 22; 23 is a flowchart for explaining normal predicted motion vector mode derivation processing of the normal predicted motion vector mode deriving unit 401 of FIG. 22; 17 is a block diagram of a detailed configuration of a sub-block predicted motion vector mode derivation unit 303 of FIG. 16. FIG. 23 is a block diagram of a detailed configuration of a sub-block predicted motion vector mode derivation unit 403 of FIG. 22. FIG. FIG. 17 is a block diagram of a detailed configuration of a sub-block merging mode derivation unit 304 of FIG. 16; 23 is a block diagram of a detailed configuration of a sub-block merging mode derivation unit 404 of FIG. 22; FIG. FIG. 10 is a diagram illustrating derivation of affine inherited motion vector predictor candidates; FIG. 4 is a diagram for explaining derivation of affine-constructed motion vector predictor candidates; FIG. 10 is a diagram illustrating affine inheritance merging candidate derivation; FIG. 10 is a diagram illustrating affine construction merging candidate derivation; FIG. 11 is a flowchart of affine inheritance motion vector predictor candidate derivation; FIG. 10 is a flowchart of deriving affine-constructed motion vector predictor candidates. 10 is a flowchart of affine inheritance merge candidate derivation. 10 is a flow chart of affine construction merge candidate derivation. FIG. 11 is a flowchart for explaining a history motion vector predictor candidate list initialization/update processing procedure; FIG. FIG. 10 is a flowchart of a same-element confirmation process procedure in the historical motion vector predictor candidate derivation process procedure; FIG. FIG. 10 is a flowchart of an element shift processing procedure in the historical motion vector predictor candidate derivation processing procedure; FIG. FIG. 10 is a flowchart for explaining a history motion vector predictor candidate derivation processing procedure; FIG. FIG. 11 is a flowchart for explaining history merge candidate derivation processing procedures; FIG. FIG. 10 is a diagram illustrating an example of history motion vector predictor candidate list update processing; 11 is a flowchart for explaining the operation of a sub-block temporal merge candidate derivation unit 381; FIG. 11 is a flowchart for explaining processing for deriving adjacent motion information of blocks; FIG. 4 is a flowchart for explaining processing for deriving a temporal motion vector; 4 is a flowchart for explaining derivation of inter prediction information; FIG. 10 is a flowchart for explaining processing for deriving sub-block motion information; FIG. FIG. 4 is a diagram for explaining the temporal anteroposterior relationship of pictures; 10 is a flowchart for explaining derivation processing of a temporal motion vector predictor candidate in a normal motion vector predictor mode derivation unit 301. FIG. 10 is a flowchart for explaining derivation processing of ColPic in derivation processing of temporal motion vector predictor candidates in the normal motion vector predictor mode derivation unit 301. FIG. 10 is a flowchart for explaining processing for deriving ColPic encoding information in derivation processing of a temporal motion vector predictor candidate in a normal motion vector predictor mode deriving unit 301. FIG. FIG. 11 is a flowchart for explaining derivation processing of inter prediction information; FIG. FIG. 10 is a flowchart showing a procedure for deriving inter prediction information of a coding block when the inter prediction mode of the coding block colCb is bi-prediction (Pred_BI); FIG. FIG. 10 is a flowchart for explaining a motion vector scaling calculation processing procedure; FIG. FIG. 11 is a flowchart for explaining derivation processing of temporal merge candidates; FIG. FIG. 10 is a diagram for explaining motion compensation prediction in the case of L0 prediction in which the reference picture (RefL0Pic) of L0 is before the picture to be processed (CurPic); FIG. 11 is a diagram for explaining motion compensation prediction when L0 prediction is performed and the reference picture for L0 prediction is at a time after the picture to be processed; FIG. 10 is a diagram for explaining motion compensation prediction in the case of bi-prediction where the reference picture for L0 prediction is before the picture to be processed and the reference picture for L1 prediction is after the picture to be processed; . FIG. 10 is a diagram for explaining the prediction direction of motion compensated prediction when bi-prediction is performed and the reference picture for L0 prediction and the reference picture for L1 prediction are located before the picture to be processed; FIG. 10 is a diagram for explaining the prediction direction of motion compensation prediction when bi-prediction is performed and the reference picture for L0 prediction and the reference picture for L1 prediction are located after the picture to be processed; FIG. 11 is a flowchart for explaining an average merge candidate derivation processing procedure; FIG. 10 is a table showing information about merge difference motion vectors; FIG. 10 is a diagram illustrating derivation of a merge difference motion vector; 10 is a flowchart for explaining affine transformation motion compensation processing; FIG. 4 is a diagram for explaining reference positions for deriving motion vectors of sub-blocks;

Techniques and technical terms used in this embodiment are defined.

<Tree Block>
In the embodiment, an image to be encoded/decoded is equally divided into a predetermined size. This unit is defined as a treeblock. As shown in FIG. 4, the size of the treeblock is set to 128×128 pixels in this embodiment, but the size of the treeblock is not limited to this, and any size may be set. The treeblock to be processed (corresponding to the encoding target in the encoding process and the decoding target in the decoding process) is switched in raster scan order, that is, from left to right and from top to bottom. The interior of each treeblock can be further recursively split. A block to be coded/decoded after treeblock division is defined as a coded block. Also, tree blocks and coding blocks are collectively defined as blocks. Appropriate block division enables efficient encoding. The size of the treeblock can be a fixed value that is prearranged between the encoding device and the decoding device, or can be configured such that the treeblock size determined by the encoding device is transmitted to the decoding device.

<Prediction mode>
For each encoding block to be processed, the image to be processed has been processed (in the encoding process, the signal that has been encoded is used for the decoded image, image signal, tree block, block, encoding block, etc. In the decoding process, Intra prediction (MODE_INTRA) that performs prediction from the surrounding image signals of decoded images, image signals, tree blocks, blocks, coding blocks, etc.), and inter prediction that performs prediction from the image signals of processed images Switch (MODE_INTER). A mode for distinguishing between intra prediction (MODE_INTRA) and inter prediction (MODE_INTER) is defined as a prediction mode (PredMode). The prediction mode (PredMode) has intra prediction (MODE_INTRA) or inter prediction (MODE_INTER) as a value.

<Inter Prediction>
In inter prediction, in which prediction is performed from image signals of processed images, a plurality of processed images can be used as reference pictures. In order to manage a plurality of reference pictures, two types of reference lists, L0 (reference list 0) and L1 (reference list 1), are defined, and reference pictures are specified using reference indices. L0 prediction (Pred_L0) is available for P slices. L0 prediction (Pred_L0), L1 prediction (Pred_L1), and bi-prediction (Pred_BI) are available for B slices. L0 prediction (Pred_L0) is inter prediction that refers to reference pictures managed by L0, and L1 prediction (Pred_L1) is inter prediction that refers to reference pictures that are managed by L1. Bi-prediction (Pred_BI) is inter-prediction in which both L0 prediction and L1 prediction are performed and one reference picture managed by each of L0 and L1 is referred to. Information specifying L0 prediction, L1 prediction, and bi-prediction is defined as an inter-prediction mode. In the subsequent processing, it is assumed that the constants and variables with the suffix LX attached to the output are processed for each of L0 and L1.

<Predicted motion vector mode>
The motion vector predictor mode is a mode in which an index for specifying a motion vector predictor, a differential motion vector, an inter prediction mode, and a reference index are transmitted, and inter prediction information of the block to be processed is determined. The motion vector predictor is a motion vector predictor candidate derived from a processed block adjacent to the target block or a block belonging to the processed image and located at the same position or in the vicinity of the target block (neighborhood), and a motion vector predictor. Derived from the index to identify the vector.

<Merge mode>
Merge mode is a processed block adjacent to the target block, or a block belonging to the processed image and located at the same position as the target block or its vicinity (neighborhood) without transmitting the differential motion vector and the reference index. This mode derives the inter prediction information of the block to be processed from the inter prediction information of the target block.

A processed block adjacent to the block to be processed and the inter prediction information of the processed block are defined as spatial merge candidates. A block belonging to a processed image and located at the same position as or in the vicinity (neighborhood) of the block to be processed and inter prediction information derived from the inter prediction information of the block are defined as temporal merge candidates. Each merging candidate is registered in a merging candidate list, and a merging candidate to be used in prediction of a block to be processed is specified by a merging index.

<Adjacent block>
FIG. 11 is a diagram illustrating reference blocks referred to for deriving inter prediction information in motion vector predictor mode and merge mode. A0, A1, A2, B0, B1, B2, and B3 are processed blocks adjacent to the target block. T0 is a block belonging to the processed image, and is a block located at the same position as or near (neighborhood to) the encoding block to be processed of the image to be processed.

A1 and A2 are blocks located on the left side of the target encoding block and adjacent to the target encoding block. B1 and B3 are blocks positioned above the target encoding block and adjacent to the target encoding block. A0, B0, and B2 are blocks positioned at the lower left, upper right, and upper left of the encoding block to be processed, respectively.

The details of how adjacent blocks are handled in motion vector predictor mode and merge mode will be described later.

<Affine transformation motion compensation>
Affine transformation motion compensation divides a coding block into sub-blocks of a predetermined unit, sets a motion vector individually for each sub-block, and performs motion compensation. The motion vector of each sub-block is derived from the inter-prediction information of a processed block adjacent to the target block or a block belonging to the processed image and located at or near the same position as the target block (neighborhood)1 Based on one or more control points. In this embodiment, the size of the sub-block is 4×4 pixels, but the size of the sub-block is not limited to this, and the motion vector may be derived in units of pixels.

FIG. 14 shows an example of affine transform motion compensation when there are two control points. In this case, since the two control points have two parameters of a horizontal component and a vertical component, the affine transformation with two control points is called a four-parameter affine transformation. CP1 and CP2 in FIG. 14 are control points. FIG. 15 shows an example of affine transform motion compensation with three control points. In this case, since the three control points have two parameters of a horizontal component and a vertical component, the affine transformation with three control points is called a six-parameter affine transformation. CP1, CP2, and CP3 in FIG. 15 are control points.

Affine transform motion compensation can be used in both predictive motion vector mode and merge mode. A mode in which affine transform motion compensation is applied in motion vector predictor mode is defined as sub-block motion vector predictor mode, and a mode in which affine transform motion compensation is applied in merge mode is defined as sub-block merge mode.

<Inter prediction syntax>
The syntax for inter prediction will be described with reference to FIGS. 12 and 13. FIG. merge_flag in FIG. 12 is a flag indicating whether the encoding block to be processed is set to merge mode or motion vector predictor mode. merge_affine_flag is a flag indicating whether or not to apply the sub-block merge mode to an encoding block to be processed in merge mode. inter_affine_flag is a flag indicating whether or not the sub-block motion vector predictor mode is applied to the encoding block to be processed in the motion vector predictor mode. cu_affine_type_flag is a flag for determining the number of control points in the sub-block predicted motion vector mode. FIG. 13 shows the value of each syntax element and the prediction method corresponding thereto. merge_flag=1,merge_affine_flag=0 corresponds to normal merge mode, which is a non-subblock merge mode. merge_flag=1,merge_affine_flag=1 corresponds to subblock merge mode. merge_flag=0, inter_affine_flag=0 corresponds to normal motion vector predictor mode, which is motion vector predictor merge that is not sub-block motion vector predictor mode. merge_flag=0, inter_affine_flag=1 corresponds to sub-block prediction motion vector mode. If merge_flag=0 and inter_affine_flag=1, then transmit cu_affine_type_flag to determine the number of control points.

<POC>
POC (Picture Order Count) is a variable associated with a picture to be encoded, and is set to a value that increases by 1 in the order in which pictures are output. Based on the POC value, it is possible to determine whether the pictures are the same, determine the sequential relationship between the pictures in the output order, and derive the distance between the pictures. For example, if two pictures have the same POC value, it can be determined that they are the same picture. When two pictures have different POC values, it can be determined that the picture with the smaller POC value is the picture to be output first, and the difference between the POCs of the two pictures determines the inter-picture distance in the time axis direction. show.

(First embodiment)
An image encoding device 100 and an image decoding device 200 according to the first embodiment of the present invention will be described.

FIG. 1 is a block diagram of an image coding device 100 according to the first embodiment. The moving image coding apparatus according to the embodiment includes an image coding apparatus 100, a block division unit 101, an inter prediction unit 102, an intra prediction unit 103, a decoded image memory 104, a prediction method determination unit 105, a residual signal generation unit 106, An orthogonal transform/quantization unit 107 , a bit stream encoding unit 108 , an inverse quantization/inverse orthogonal transform unit 109 , a decoded image signal superimposing unit 110 , and an encoded information storage memory 111 are provided.

The block division unit 101 recursively divides an input image to generate coding blocks. The block dividing unit 101 includes a 4-dividing unit that horizontally and vertically divides a block to be divided, and a 2-3 dividing unit that horizontally or vertically divides a block to be divided. . The generated image signal of the coding block to be processed is supplied to the inter prediction unit 102 , the intra prediction unit 103 and the residual signal generation unit 106 . Information indicating the determined recursive partitioning structure is also supplied to the bit stream encoding unit 108 . A detailed operation of the block division unit 101 will be described later.

The inter prediction unit 102 performs inter prediction of the coding block to be processed. A plurality of inter prediction information candidates are derived from the inter prediction information stored in the encoded information storage memory and the decoded image signal stored in the decoded image memory 104, and suitable inter prediction is performed from among the plurality of candidates. A mode is selected, and the selected inter-prediction mode and the predicted image signal corresponding to the selected inter-prediction mode are supplied to the prediction method determination unit 105 . A detailed configuration and operation of the inter prediction unit 102 will be described later.

The intra prediction unit 103 performs intra prediction on the encoding block to be processed. Generate a predicted image signal by intra prediction from the decoded image signal stored in the decoded image memory 104, select a suitable intra prediction mode from among a plurality of intra prediction modes, select the selected intra prediction mode, and A predicted image signal corresponding to the selected intra prediction mode is supplied to the prediction method determination unit 105 . FIG. 10 shows an example of intra prediction. FIG. 10A shows the correspondence between prediction directions of intra prediction and intra prediction mode numbers. For example, intra-prediction mode 50 produces intra-predicted images by copying pixels vertically. Intra-prediction mode 1 is a DC mode, and is a mode in which all pixel values of a block to be processed are average values of reference pixels. Intra prediction mode 0 is a planar mode, and is a mode in which a two-dimensional intra prediction image is created from reference pixels in the vertical and horizontal directions. FIG. 10B is an example of generating an intra-prediction image in intra-prediction mode 40. FIG. The value of the reference pixel in the direction indicated by the intra prediction mode is copied to each pixel of the block to be processed. When the reference pixel in the intra prediction mode is not at the integer position, the reference pixel value is determined by interpolation from the reference pixel values at the surrounding integer positions.

The decoded image memory 104 stores the decoded image generated by the decoded image signal superimposing unit 110 . The decoded image stored in the decoded image memory is supplied to the inter prediction section 102 and intra prediction section 103 .

The prediction method determination unit 105 evaluates each prediction using the encoded information and the code amount of the residual signal, the amount of distortion between the predicted image signal and the image signal to be processed, and the like, thereby determining the optimum prediction. Determine the mode (inter prediction or intra prediction). In the case of the inter-prediction merge mode, the coded information of the merge index and the information indicating whether the sub-block merge mode is selected (sub-block merge flag) is supplied to the bit stream coding unit 108, and the inter-prediction motion vector prediction mode is selected. In this case, bitstream encoding section 108 encodes encoded information such as inter prediction mode, predicted motion vector index, L0 and L1 reference indices, differential motion vector, information indicating whether sub-block mode (sub-block predicted motion vector flag), etc. supply to The determined encoded information is supplied to the encoded information storage memory 111 .

The residual signal generation unit 106 generates a residual signal by subtracting the predicted image signal from the image signal to be processed, and supplies the residual signal to the orthogonal transformation/quantization unit 107 .

Orthogonal transformation/quantization section 107 performs orthogonal transformation and quantization on the residual signal according to a quantization parameter to generate an orthogonal transformation/quantized residual signal, and performs bit stream encoding section 108 and inverse quantization. supplied to the normalization/inverse orthogonal transformation unit 109 .

The bit stream encoding unit 108 encodes encoding information according to the prediction method determined by the prediction method determination unit 105 for each encoding block, in addition to the information for each sequence, picture, slice, and encoding block. Specifically, in the case of prediction mode PredMode, partition mode PartMode, and inter prediction (PRED_INTER) for each coding block, a flag to determine whether it is merge mode, sub-block merge flag, merge index in case of merge mode, merge If it is not the mode, the encoding information such as the inter prediction mode, the predicted motion vector index, the difference motion vector information, the sub-block predicted motion vector flag, etc. is encoded according to the prescribed syntax rule described later to generate the first encoded bit stream. do. Further, bit string encoding section 108 entropy-encodes the orthogonally transformed and quantized residual signal according to a prescribed syntax rule to generate a second encoded bit string. The first coded bit stream and the second coded bit stream are multiplexed according to a prescribed syntax rule to output a bit stream.

The inverse quantization/inverse orthogonal transformation unit 109 performs inverse quantization and inverse orthogonal transformation on the orthogonally transformed/quantized residual signal supplied from the orthogonal transformation/quantization unit 107 to calculate a residual signal, and decodes it. It is supplied to the image signal superimposing unit 110 .

The decoded image signal superimposition unit 110 superimposes the prediction image signal determined by the prediction method determination unit 105 and the residual signal inversely quantized and inverse orthogonally transformed by the inverse quantization/inverse orthogonal transformation unit 109 to generate a decoded image. is generated and stored in the decoded image memory 104 . Note that the decoded image may be stored in the decoded image memory 104 after being subjected to filtering processing for reducing distortion such as block distortion due to encoding.

The coding information storage memory 111 stores coding information such as the prediction mode (inter prediction or intra prediction) determined by the prediction method determination unit 105 . The coding information stored in the coding information storage memory 111 includes, in the case of inter prediction, the determined motion vector, reference list, and reference index; Coding information of information indicating whether or not (sub-block merge flag), inter prediction mode in case of motion vector predictor mode of inter prediction, motion vector predictor index, reference index of L0 and L1, differential motion vector, sub-block mode Information indicating whether or not (sub-block prediction motion vector flag), in the case of intra prediction, the determined intra prediction mode, and the like. Construction of the history candidate list managed by the encoded information storage memory 111 will be described later.

FIG. 2 is a block diagram showing the configuration of a video decoding device according to an embodiment of the present invention corresponding to the video encoding device of FIG. The video decoding device of the embodiment includes a bit string decoding unit 201, a block division unit 202, an inter prediction unit 203, an intra prediction unit 204, an encoded information storage memory 205, an inverse quantization/inverse orthogonal transformation unit 206, a decoded image signal A superimposing unit 207 and a decoded image memory 208 are provided.

The decoding process of the video decoding device in FIG. 2 corresponds to the decoding process provided inside the video encoding device in FIG. The components of the inverse orthogonal transform unit 206, the decoded image signal superimposition unit 207, and the decoded image memory 208 are the inverse quantization/inverse orthogonal transform unit 109, the decoded image signal superimposition unit 110, and the It has a function corresponding to each configuration of the encoded information storage memory 111 and the decoded image memory 104 .

The bitstream supplied to the bitstream decoding unit 201 is separated according to a prescribed syntax rule. The separated first encoded bit stream is decoded to obtain sequences, pictures, slices, information in units of encoded blocks, and encoded information in units of encoded blocks. Specifically, the prediction mode PredMode that distinguishes between inter prediction (PRED_INTER) and intra prediction (PRED_INTRA) for each coding block, the partition mode PartMode, and in the case of inter prediction (PRED_INTER), a flag that distinguishes whether or not it is merge mode , merge index, sub-block merge flag for merge mode, inter-prediction mode for motion vector predictor mode, motion vector predictor index, differential motion vector, sub-block predictor motion vector flag, etc. and supplies the encoded information to the inter prediction unit 203 or intra prediction unit 204 and the encoded information storage memory 205 . The separated second encoded bit string is decoded to calculate an orthogonally transformed and quantized residual signal, and the orthogonally transformed and quantized residual signal is supplied to the inverse quantization and inverse orthogonal transformation section 206 .

When the prediction mode PredMode of the encoding block to be processed is inter prediction (PRED_INTER) and motion vector prediction mode, the inter prediction unit 203 predicts the code of the already decoded image signal stored in the encoding information storage memory 205. Using the transformation information, a plurality of motion vector predictor candidates are derived and registered in a motion vector predictor candidate list, which will be described later. A motion vector predictor corresponding to a motion vector predictor index decoded and supplied by the unit 201 is selected, a motion vector is calculated from the differential vector decoded by the bit string decoding unit 201 and the selected motion vector predictor, and other encoding is performed. It is stored in the encoded information storage memory 205 together with the information. The coding information of the coding block supplied and stored here includes flags predFlagL0[xP][yP], predFlagL1[xP] indicating whether or not to use prediction mode PredMode, partition mode PartMode, L0 prediction, and L1 prediction. [yP], reference indices refIdxL0[xP][yP], refIdxL1[xP][yP] of L0, L1, motion vectors mvL0[xP][yP], mvL1[xP][yP] of L0, L1, etc. . Here, xP and yP are indices indicating the position of the upper left pixel of the coding block in the picture. When the prediction mode PredMode is inter prediction (MODE_INTER) and the inter prediction mode is L0 prediction (Pred_L0), the flag predFlagL0 indicating whether to use L0 prediction is 1, and the flag predFlagL1 indicates whether to use L1 prediction. is 0. When the inter prediction mode is L1 prediction (Pred_L1), the flag predFlagL0 indicating whether to use L0 prediction is 0, and the flag predFlagL1 indicating whether to use L1 prediction is 1. When the inter prediction mode is bi-prediction (Pred_BI), the flag predFlagL0 indicating whether to use L0 prediction and the flag predFlagL1 indicating whether to use L1 prediction are both 1. Furthermore, when the prediction mode PredMode of the encoding block to be processed is inter prediction (PRED_INTER) and merge mode, merge candidates are derived. A plurality of merging candidates are derived using the coding information of already decoded coding blocks stored in the coding information storage memory 205 and registered in a merge candidate list, which will be described later. A merge candidate corresponding to the merge index decoded and supplied by the bit string decoding unit 201 is selected from among the plurality of merge candidates, and a flag indicating whether to use the L0 prediction and L1 prediction of the selected merge candidate. predFlagL0[xP][yP], predFlagL1[xP][yP], L0, L1 reference index refIdxL0[xP][yP], refIdxL1[xP][yP], L0, L1 motion vector mvL0[xP][yP ], mvL1[xP][yP], etc. are stored in the encoded information storage memory 205 . Here, xP and yP are indices indicating the position of the upper left pixel of the coding block in the picture. The detailed configuration and operation of the inter prediction unit will be described later.

The intra prediction unit 204 performs intra prediction when the prediction mode PredMode of the encoding block to be processed is intra prediction (PRED_INTRA). The encoded information decoded by the bit string decoding unit 201 includes an intra prediction mode, and a predicted image signal is obtained by intra prediction from the decoded image signal stored in the decoded image memory 208 according to the intra prediction mode. is generated, and the predicted image signal is supplied to the decoded image signal superimposing unit 207 . Since the intra prediction unit 204 corresponds to the intra prediction unit 103 of the image encoding device 100, the same processing as the intra prediction unit 103 is performed.

The inverse quantization/inverse orthogonal transformation unit 206 performs inverse orthogonal transformation and inverse quantization on the orthogonally transformed/quantized residual signal decoded by the bit stream decoding unit 201, and performs inverse orthogonal transformation/inverse quantization. to obtain the residual signal.

The decoded image signal superimposing unit 207 performs inverse orthogonal transform/transform/ The decoded image signal is decoded by superimposing it with the inverse quantized residual signal and stored in the decoded image memory 208 . When storing the decoded image in the decoded image memory 208, the decoded image may be stored in the decoded image memory 208 after being subjected to a filtering process for reducing block distortion due to encoding.

Next, the operation of the block dividing section 101 in the image encoding device 100 will be described. FIG. 3 is a flowchart illustrating the operation of dividing an image into treeblocks and subdividing each treeblock. First, an input image is divided into tree blocks of a predetermined size (step S1001). Each treeblock is scanned in a predetermined order, that is, in raster scan order (step S1002), and the inside of the treeblock to be processed is divided (step S1003).

FIG. 7 is a flow chart showing the detailed operation of the division processing in step S1003. First, it is determined whether or not to divide the block to be processed into four (step S1101).

If it is determined that the block to be processed is to be divided into four, the block to be processed is divided into four (step S1102). Each block obtained by dividing the block to be processed is scanned in the order of Z scanning, that is, upper left, upper right, lower left, and lower right (step S1103). FIG. 5 shows an example of the Z scan order, and 601 in FIG. 6 shows an example in which the block to be processed is divided into four. Numbers 0 to 3 in 601 in FIG. 6 indicate the order of processing. Then, the flowchart of FIG. 7 is recursively called for each block divided in step S1101.

If it is determined that the block to be processed is not divided into 4, it is divided into 2-3 (step S1105).

FIG. 8 is a flow chart showing the detailed operation of the 2-3 division processing in step S1105. First, it is determined whether or not to divide the block to be processed into 2-3, that is, whether to divide into 2 or 3 (step S1201).

If it is determined not to divide the block to be processed into 2-3, that is, if it is determined not to be divided, the division is terminated (step S1211), and the block in the upper layer is returned to.

If it is determined to divide the block to be processed into 2 or 3, it is determined whether to further divide the block to be processed into two (step S1202).

If the block to be processed is determined to be divided into two, it is determined whether to divide the block to be processed vertically (step S1203), and based on the result, the block to be processed is vertically divided (step S1204). Alternatively, the block to be processed is horizontally divided (step S1205). As a result of step S1204, the block to be processed is divided into two vertically as shown in FIG. 602, and as a result of step S1205, the block to be processed is divided into two horizontally as shown in FIG.

If it is determined in step S1202 that the block to be processed is not to be divided into two, that is, if it is determined to be divided into three, it is determined whether or not to divide the block to be processed in the vertical direction (step S1206). Based on this, the block to be processed is divided vertically (step S1207) or horizontally (step S1208). As a result of step S1207, the block to be processed is divided vertically into three parts as shown in FIG. 603, and as a result of step S1208, the block to be processed is divided into three parts horizontally as shown in FIG.

After executing any one of steps S1204 to S1205, each block obtained by dividing the block to be processed is scanned from left to right and from top to bottom (step S1209). Numbers 0 to 3 in 602 to 605 in FIG. 6 indicate the order of processing. The flowchart of FIG. 8 is recursively called for each divided block.

In the recursive block division described here, necessity of division may be restricted by the number of divisions, the size of the block to be processed, or the like. Information that limits whether or not division is necessary may be implemented by a configuration in which information is not transmitted by prior agreement between the encoding device and the decoding device, or the coding device may limit whether or not division is necessary. The information may be determined and recorded in the encoded bit string to be transmitted to the decoding device.

Here, when a certain block is divided, the block before division is called a parent block, and each block after division is called a child block.

Next, the operation of the block dividing section 202 in the image decoding device 200 will be described. The block dividing unit 202 divides the treeblocks in the same processing procedure as the block dividing unit 101 of the image encoding device 100 . However, the block division unit 101 of the image encoding device 100 applies optimization techniques such as estimation of the optimum shape by image recognition and distortion rate optimization to determine the optimum block division shape, whereas the image decoding device The block division unit 202 in 200 is different in that the block division shape is determined by decoding the block division information recorded in the encoded bit stream.

FIG. 9 shows the syntax (syntax rules for encoded bit strings) relating to block division in the first embodiment. coding_quadtree() represents the syntax for dividing a block into four, and multi_type_tree() represents the syntax for dividing the block into two or three. qt_split is a flag indicating whether to divide the block into four or not. When the block is divided into four, qt_split=1, and when not divided into four, qt_split=0. When dividing into four (qt_split=1), each block divided into four is recursively divided into four (coding_quadtree(0), coding_quadtree(1), coding_quadtree(2), coding_quadtree(3)). When not splitting into 4 (qt_split=0), the subsequent splitting is determined according to multi_type_tree(). mtt_split is a flag indicating whether or not to further split. For further splitting (mtt_split=1), refer to mtt_split_vertical, which is a flag indicating whether to split vertically or horizontally, and mtt_split_binary, which is a flag for determining whether to split into two or three. mtt_split_vertical=1 indicates splitting in the vertical direction, and mtt_split_vertical=0 indicates splitting in the horizontal direction. mtt_split_binary=1 indicates splitting into two, and mtt_split_binary=0 indicates splitting into three. Perform hierarchical block splitting by recursively calling multi_type_tree until mtt_split=0.

<Inter Prediction>
The inter prediction method according to the embodiment is implemented in the inter prediction section 102 of the video encoding device in FIG. 1 and the inter prediction section 203 of the video decoding device in FIG.

An inter-prediction method according to an embodiment will be described with reference to the drawings. The inter-prediction method is implemented in both encoding and decoding processes in units of encoded blocks.

<Description of inter prediction unit 102 on the encoding side>
FIG. 16 is a diagram showing the detailed configuration of the inter prediction unit 102 of the video encoding device of FIG. A normal motion vector predictor mode deriving unit 301 derives a plurality of normal motion vector predictor candidates, selects a motion vector predictor, and calculates a differential vector from the detected motion vector. The detected inter prediction mode, reference index, motion vector, and calculated difference vector are inter prediction information for the normal predicted motion vector mode. This inter prediction information is supplied to the inter prediction mode determination unit 305 . The detailed configuration and processing of the normal motion vector predictor mode derivation unit 301 will be described later.

A normal merge mode derivation unit 302 derives a plurality of normal merge candidates, selects a normal merge candidate, and obtains inter prediction information in the normal merge mode. This inter prediction information is supplied to the inter prediction mode determination unit 305 . The detailed configuration and processing of the normal merge mode derivation unit 302 will be described later.

A sub-block predictor motion vector mode deriving unit 303 derives a plurality of sub-block predictor motion vector candidates, selects a sub-block predictor motion vector, and calculates a difference vector from the detected motion vector. The detected inter prediction mode, reference index, motion vector, and calculated difference vector are inter prediction information for the normal predicted motion vector mode. This inter prediction information is supplied to the inter prediction mode determination unit 305 . The detailed configuration and processing of the sub-block predicted motion vector mode derivation unit 303 will be described later.

A sub-block merging mode derivation unit 304 derives a plurality of sub-block merging candidates, selects a sub-block merging candidate, and obtains inter-prediction information for the sub-block merging mode. This inter prediction information is supplied to the inter prediction mode determination unit 305 . A detailed configuration and processing of the sub-block merge mode derivation unit 304 will be described later.

Based on the inter prediction information supplied from the normal prediction motion vector mode derivation unit 301, the normal merge mode derivation unit 302, the sub-block prediction motion vector mode derivation unit 303, and the sub-block merge mode derivation unit 304, the inter prediction mode determination unit 305 , determine the inter-prediction mode. The inter prediction information according to the determination result is supplied from the inter prediction mode determination unit 305 to the motion compensation prediction unit 306 .

Based on the determined inter prediction information, the motion compensation prediction unit 306 performs inter prediction on the reference image signal stored in the decoded image memory 104 . Detailed configuration and processing will be described later.

<Description of Inter Predictor 203 on the Decoding Side>
FIG. 22 is a diagram showing a detailed configuration of the inter prediction unit 203 of the video decoding device of FIG. 2. As shown in FIG.

A normal motion vector predictor mode deriving unit 401 derives a plurality of normal motion vector predictor candidates, selects a motion vector predictor, and calculates a differential vector from the detected motion vector. The detected inter prediction mode, reference index, motion vector, and difference vector are inter prediction information for the normal predicted motion vector mode. This inter prediction information is supplied to the motion compensation prediction section 406 via the switch 408 . The detailed configuration and processing of the normal motion vector predictor mode derivation unit 401 will be described later.

A normal merge mode derivation unit 402 derives a plurality of normal merge candidates, selects a normal merge candidate, and obtains inter prediction information in the normal merge mode. This inter prediction information is supplied to the motion compensation prediction section 406 via the switch 408 . The detailed configuration and processing of the normal merge mode derivation unit 402 will be described later.

A sub-block predictor motion vector mode derivation unit 403 derives a plurality of sub-block predictor motion vector candidates, selects a sub-block predictor motion vector, and calculates a difference vector from the detected motion vector. The detected inter prediction mode, reference index, motion vector, and calculated difference vector are inter prediction information for the normal predicted motion vector mode. This inter prediction information is supplied to the motion compensation prediction section 406 via the switch 408 . The detailed configuration and processing of the sub-block predicted motion vector mode deriving unit 403 will be described later.

A sub-block merging mode deriving unit 404 derives a plurality of sub-block merging candidates, selects a sub-block merging candidate, and obtains inter-prediction information for the sub-block merging mode. This inter prediction information is supplied to the motion compensation prediction section 406 via the switch 408 . A detailed configuration and processing of the sub-block merge mode derivation unit 404 will be described later.

Based on the determined inter prediction information, the motion compensation prediction unit 406 performs inter prediction on the reference image signal stored in the decoded image memory 208 . The detailed configuration and processing are the same as those on the encoding side.

<Normal predicted motion vector mode derivation unit (normal AMVP)>
The normal motion vector predictor mode deriving unit 301 in FIG. A vector detection unit 326 , a motion vector predictor candidate selection unit 327 and a motion vector subtraction unit 328 are included.

The normal motion vector predictor mode deriving unit 401 in FIG. A vector candidate selection unit 426 and a motion vector addition unit 427 are included.

Processing procedures of the normal predicted motion vector mode deriving unit 301 on the encoding side and the normal predicted motion vector mode deriving unit 401 on the decoding side will be described with reference to the flowcharts of FIGS. 19 and 25, respectively. FIG. 19 is a flowchart showing the normal predicted motion vector mode derivation processing procedure by the normal motion vector mode derivation unit 301 on the encoding side, and FIG. 25 is the normal predicted motion vector mode derivation processing by the normal motion vector mode derivation unit 401 on the decoding side. It is a flow chart which shows a procedure.

<Normal predicted motion vector mode derivation unit (normal AMVP): description on the encoding side>
A normal motion vector prediction mode derivation processing procedure on the encoding side will be described with reference to FIG. In the explanation of the processing procedure in FIG. 19, the term "motion vector" in the specification corresponds to the term "normal motion vector" in FIG. First, the normal motion vector detection unit 326 detects a normal motion vector for each inter prediction mode and reference index (step S100 in FIG. 19).

Subsequently, a spatial motion vector predictor candidate deriving unit 321, a temporal motion vector predictor candidate deriving unit 322, a historical motion vector predictor candidate deriving unit 323, a motion vector predictor candidate supplementing unit 325, a motion vector predictor candidate selecting unit 327, and a motion vector subtracting unit. At 328, differential motion vectors of motion vectors used for inter prediction in normal motion vector predictor mode are calculated for each of L0 and L1 (steps S101 to S106 in FIG. 19). Specifically, when the prediction mode PredMode of the block to be processed is inter prediction (MODE_INTER) and the inter prediction mode is L0 prediction (Pred_L0), the L0 motion vector predictor candidate list mvpListL0 is calculated and the motion vector predictor mvpL0 is selected. Then, the difference motion vector mvdL0 of the motion vector mvL0 of L0 is calculated. When the inter prediction mode of the block to be processed is L1 prediction (Pred_L1), the L1 motion vector predictor candidate list mvpListL1 is calculated, the motion vector predictor mvpL1 is selected, and the differential motion vector mvdL1 of the L1 motion vector mvL1 is calculated. . When the inter prediction mode of the block to be processed is bi-prediction (Pred_BI), both L0 prediction and L1 prediction are performed, L0 motion vector predictor candidate list mvpListL0 is calculated, L0 motion vector predictor mvpL0 is selected, and L0 Calculates the differential motion vector mvdL0 of the motion vector mvL0 of L1, calculates the motion vector predictor candidate list mvpListL1 of L1, calculates the predicted motion vector mvpL1 of L1, and calculates the differential motion vector mvdL1 of the motion vector mvL1 of L1. do.

Differential motion vector calculation processing is performed for each of L0 and L1, and the processing is common to both L0 and L1. Therefore, L0 and L1 are represented as common LX in the following description. X is 0 in the process of calculating the differential motion vector of L0, and X is 1 in the process of calculating the differential motion vector of L1. Also, when referring to the information of the other list instead of LX during the process of calculating the differential motion vector of LX, the other list is represented as LY.

If the LX motion vector mvLX is used (step S102 in FIG. 19: YES), the LX motion vector predictor candidates are calculated and the LX motion vector predictor candidate list mvpListLX is constructed (step S103 in FIG. 19). A spatial motion vector predictor candidate deriving unit 321, a temporal motion vector predictor candidate deriving unit 322, a historical motion vector predictor candidate deriving unit 323, and a motion vector predictor candidate supplementing unit 325 in the normal motion vector predictor mode deriving unit 301 generate multiple predicted motions. A motion vector predictor candidate list mvpListLX is constructed by deriving vector candidates. A detailed processing procedure of step S103 in FIG. 19 will be described later using the flowchart in FIG.

Subsequently, the motion vector predictor candidate selection unit 327 selects the LX motion vector predictor mvpLX from the LX motion vector predictor candidate list mvpListLX (step S104 in FIG. 19). A motion vector difference between the motion vector mvLX and each motion vector predictor candidate mvpListLX[i] stored in the motion vector predictor candidate list mvpListLX is calculated. A code amount when the differential motion vectors are encoded is calculated for each element of the motion vector predictor candidate list mvpListLX. Then, among the elements registered in the motion vector predictor candidate list mvpListLX, the motion vector predictor candidate mvpListLX[i] having the minimum code amount for each motion vector predictor candidate is selected as the motion vector predictor mvpLX. Get index i. If there are multiple motion vector predictor candidates with the smallest generated code amount in the motion vector predictor candidate list mvpListLX, the motion vector predictor candidate list mvpListLX with a small index i is indicated by a small number. , is selected as the optimal motion vector predictor mvpLX, and its index i is obtained.

Subsequently, the motion vector subtraction unit 328 subtracts the selected predicted motion vector mvpLX of LX from the motion vector mvLX of LX,
mvdLX = mvLX - mvpLX
, the differential motion vector mvdLX of LX is calculated (step S105 in FIG. 19).

<Normal predicted motion vector mode derivation unit (normal AMVP): description on the decoding side>
Next, the decoding-side normal predicted motion vector mode processing procedure will be described with reference to FIG. On the decoding side, the spatial motion vector predictor candidate deriving unit 421, the temporal motion vector predictor candidate deriving unit 422, the historical motion vector predictor candidate deriving unit 423, and the motion vector predictor candidate supplementing unit 425 use inter prediction in normal motion vector predictor mode. A motion vector is calculated for each of L0 and L1 (steps S201 to S206 in FIG. 25). Specifically, when the prediction mode PredMode of the block to be processed is inter prediction (MODE_INTER) and the inter prediction mode of the block to be processed is L0 prediction (Pred_L0), the L0 motion vector predictor candidate list mvpListL0 is calculated and the predicted motion Select the vector mvpL0 and calculate the motion vector mvL0 of L0. When the inter prediction mode of the block to be processed is L1 prediction (Pred_L1), the L1 motion vector predictor candidate list mvpListL1 is calculated, the motion vector predictor mvpL1 is selected, and the L1 motion vector mvL1 is calculated. When the inter prediction mode of the block to be processed is bi-prediction (Pred_BI), both L0 prediction and L1 prediction are performed, L0 motion vector predictor candidate list mvpListL0 is calculated, L0 motion vector predictor mvpL0 is selected, and L0 L1's motion vector mvL0 is calculated, L1's motion vector predictor candidate list mvpListL1 is calculated, L1's predicted motion vector mvpL1 is calculated, and L1's motion vector mvL1 is calculated.

Similar to the encoding side, the decoding side also performs motion vector calculation processing for each of L0 and L1, but the processing is common to both L0 and L1. Therefore, L0 and L1 are represented as common LX in the following description. LX represents an inter-prediction mode used for inter-prediction of a coding block to be processed. X is 0 in the process of calculating the motion vector of L0, and X is 1 in the process of calculating the motion vector of L1. Also, when referring to the information of another reference list instead of the same reference list as that of LX to be calculated during the process of calculating the motion vector of LX, the other reference list is represented as LY.

If the LX motion vector mvLX is to be used (step S202 in FIG. 25: YES), LX motion vector predictor candidates are calculated to construct the LX motion vector predictor candidate list mvpListLX (step S203 in FIG. 25). A spatial motion vector predictor candidate deriving unit 421, a temporal motion vector predictor candidate deriving unit 422, a historical motion vector predictor candidate deriving unit 423, and a motion vector predictor candidate supplementing unit 425 in the normal motion vector predictor mode deriving unit 401 generate a plurality of predicted motions. Vector candidates are calculated and a motion vector predictor candidate list mvpListLX is constructed. A detailed processing procedure of step S203 in FIG. 25 will be described later using the flowchart in FIG.

Subsequently, the motion vector predictor candidate selection unit 426 selects a motion vector predictor candidate mvpListLX[mvpIdxLX] corresponding to the motion vector predictor index mvpIdxLX decoded and supplied by the bit string decoding unit 201 from the motion vector predictor candidate list mvpListLX. fetched as the predicted motion vector mvpLX (step S204 in FIG. 25).

Subsequently, the motion vector addition unit 427 adds the differential motion vector mvdLX of LX and the predicted motion vector mvpLX of LX, which are decoded and supplied by the bit string decoding unit 201, and
mvLX = mvpLX + mvdLX
, the motion vector mvLX of LX is calculated (step S205 in FIG. 25).

<Normal Predictive Motion Vector Mode Derivation Unit (Normal AMVP): Motion Vector Prediction Method>
FIG. 20 shows a normal predicted motion vector having functions common to the normal predicted motion vector mode derivation unit 301 of the video encoding device and the normal predicted motion vector mode derivation unit 401 of the video decoding device according to the embodiment of the present invention. 4 is a flow chart showing a processing procedure of mode derivation processing;

The normal motion vector predictor mode deriving unit 301 and the normal motion vector predictor mode deriving unit 401 have a motion vector predictor candidate list mvpListLX. The motion vector predictor candidate list mvpListLX has a list structure, and is provided with a motion vector predictor index indicating the location in the motion vector predictor candidate list and a storage area for storing the motion vector predictor candidate corresponding to the index as an element. The number of the motion vector predictor index starts from 0, and the motion vector predictor candidates are stored in the storage area of the motion vector predictor candidate list mvpListLX. In this embodiment, the motion vector predictor candidate list mvpListLX can register at least two motion vector predictor candidates (inter prediction information). Furthermore, 0 is set to a variable numCurrMvpCand indicating the number of motion vector predictor candidates registered in the motion vector predictor candidate list mvpListLX.

Spatial motion vector predictor candidate deriving units 321 and 421 derive a motion vector predictor candidate from the left adjacent block. In this processing, a flag availableFlagLXA indicating whether or not the motion vector predictor candidate of the left-adjacent block (A0 or A1) is available, a motion vector mvLXA, and a reference index refIdxA are derived, and mvLXA is added to the motion vector predictor candidate list mvpListLX. (step S301 in FIG. 20). Note that X is 0 when L0 and X is 1 when L1 (same below). Subsequently, the spatial motion vector predictor candidate deriving units 321 and 421 derive a motion vector predictor candidate from the upper adjacent block (B0, B1 or B2). In this process, a flag availableFlagLXB indicating whether or not the motion vector predictor candidate of the block adjacent to the upper side is available, a motion vector mvLXB, and a reference index refIdxB are derived. Add to the candidate list mvpListLX (step S302 in FIG. 20). The processes of steps S301 and S302 in FIG. 20 are common except that the positions and numbers of adjacent blocks to be referred to are different, and flag availableFlagLXN indicating whether or not the motion vector predictor candidate of the coding block is available, and motion vector mvLXN. , a reference index refIdxN (N is A or B, and so on).

Subsequently, the temporal motion vector predictor candidate deriving units 322 and 422 derive candidate motion vector predictors from the encoding blocks in the pictures whose time is different from that of the current picture to be processed. In this processing, a flag availableFlagLXCol, which indicates whether or not a motion vector predictor candidate for a coded block in a picture at a different time is available, a motion vector mvLXCol, a reference index refIdxCol, and a reference list listCol are derived, and mvLXCol is derived as a motion vector predictor candidate. Add to list mvpListLX (step S303 in FIG. 20). The derivation processing procedure in step S303 will be described later in detail.

It is assumed that the processing of the temporal motion vector predictor candidate deriving units 322 and 422 can be omitted for each sequence (SPS), picture (PPS), or slice.

Subsequently, the historical motion vector predictor candidate deriving units 323 and 423 add the historical motion vector predictor candidates registered in the historical motion vector predictor candidate list HmvpCandList to the motion vector predictor candidate list mvpListLX. (Step S304 in FIG. 20). The registration processing procedure in step S304 will be described later in detail using the flowchart of FIG.

Subsequently, the motion vector predictor candidate supplementing units 325 and 425 add motion vectors of predetermined values such as (0, 0) until the motion vector predictor candidate list mvpListLX is filled (S305 in FIG. 20).

<Normal merge mode derivation unit (normal merge)>
The normal merge mode derivation unit 302 in FIG. including.

The normal merge mode derivation unit 402 in FIG. 24 includes a spatial merge candidate derivation unit 441, a temporal merge candidate derivation unit 442, an average merge candidate derivation unit 444, a history merge candidate derivation unit 445, a merge candidate supplement unit 446, and a merge candidate selection unit 447. including.

FIG. 21 shows the procedure of normal merge mode derivation processing having functions common to the normal merge mode derivation unit 302 of the video encoding device and the normal merge mode derivation unit 402 of the video decoding device according to the embodiment of the present invention. It is a flow chart explaining.

The various processes will be described below in order. In the following description, unless otherwise specified, the slice type slice_type is B slice, but it can also be applied to P slice. However, when the slice type slice_type is a P slice, there is only L0 prediction (Pred_L0) as an inter prediction mode, and there is no L1 prediction (Pred_L1) or bi-prediction (Pred_BI), so processing related to L1 can be omitted.

The normal merge mode derivation unit 302 and the normal merge mode derivation unit 402 have a merge candidate list mergeCandList. The merge candidate list mergeCandList has a list structure, and is provided with a merge index indicating the location in the merge candidate list and a storage area for storing the merge candidate corresponding to the index as an element. The number of the merge index starts from 0, and the merge candidates are stored in the storage area of the merge candidate list mergeCandList. In the subsequent processing, merge candidates with merge index i registered in the merge candidate list mergeCandList are represented by mergeCandList[i]. In this embodiment, the merge candidate list mergeCandList can register at least six merge candidates (inter prediction information). Furthermore, 0 is set to the variable numCurrMergeCand indicating the number of merge candidates registered in the merge candidate list mergeCandList.

The spatial merging candidate deriving unit 341 and the spatial merging candidate deriving unit 441 perform processing based on the encoding information stored in the encoding information storage memory 111 of the video encoding device or the encoding information storage memory 205 of the video decoding device. Spatial merge candidates A and B are derived from the blocks adjacent to the left and upper sides of the target block, and the derived spatial merge candidates are registered in the merge candidate list mergeCandList (step S401 in FIG. 21). Now define N to denote either the spatial merge candidates A, B or the temporal merge candidates Col. Flag availableFlagN indicating whether or not inter prediction information of block N can be used as spatial merge candidate N, L0 reference index refIdxL0N and L1 reference index refIdxL1N of spatial merge candidate N, L0 indicating whether or not L0 prediction is performed A prediction flag predFlagL0N, an L1 prediction flag predFlagL1N indicating whether L1 prediction is performed, a motion vector mvL0N of L0, and a motion vector mvL1N of L1 are derived. However, in the present embodiment, merging candidates are derived without referring to other encoding blocks included in the block including the encoding block to be processed. It does not derive spatial merge candidates that

Subsequently, the temporal merge candidate derivation unit 342 and the temporal merge candidate derivation unit 442 derive temporal merge candidates from pictures at different times, and register the derived temporal merge candidates in the merge candidate list mergeCandList (see FIG. 21). step S402). A flag availableFlagCol indicating whether the temporal merge candidate is available, an L0 prediction flag predFlagL0Col indicating whether L0 prediction of the temporal merge candidate is performed, and an L1 prediction flag predFlagL1Col indicating whether L1 prediction is performed, and L0 motion vector mvL0Col of L1 and motion vector mvL1Col of L1 are derived. A detailed processing procedure of step S402 will be described later in detail with reference to FIG.

It is assumed that the processing of the temporal merge candidate derivation unit 342 and the temporal merge candidate derivation unit 442 can be omitted for each sequence (SPS), picture (PPS), or slice.

Subsequently, the history merge candidate derivation unit 345 and the history merge candidate derivation unit 445 add the history motion vector predictor candidates registered in the history motion vector predictor candidate list HmvpCandList to the merge candidate list mergeCandList (step S403 in FIG. 21). . A detailed processing procedure of step S403 will be described in detail later using the flowchart of FIG.

Subsequently, the average merge candidate derivation unit 344 and the average merge candidate derivation unit 444 derive average merge candidates from the merge candidate list mergeCandList and register the derived average merge candidates in the merge candidate list mergeCandList (step in FIG. 21). S404). A detailed processing procedure of step S404 will be described in detail later using the flowchart of FIG.

Subsequently, in the merge candidate supplementing unit 346 and the merge candidate supplementing unit 446, when the number of merge candidates numCurrMergeCand registered in the merge candidate list mergeCandList is smaller than the maximum number of merge candidates MaxNumMergeCand, the merge candidates are not registered in the merge candidate list mergeCandList. The existing merge candidate number numCurrMergeCand derives additional merge candidates up to the maximum merge candidate number MaxNumMergeCand and registers them in the merge candidate list mergeCandList (step S405 in FIG. 21). With the maximum number of merge candidates MaxNumMergeCand as the upper limit, in P slices, merge candidates whose prediction mode is L0 prediction (Pred_L0) and whose motion vector has a value of (0, 0) with different reference indices are added. In the B slice, a merging candidate whose prediction mode is bi-prediction (Pred_BI) and whose motion vector has a value of (0, 0) with a different reference index is added.

Subsequently, the merge candidate selection unit 347 and the merge candidate selection unit 447 select merge candidates from the merge candidates registered in the merge candidate list mergeCandList. The merge candidate selection unit 347 on the encoding side selects a merge candidate by calculating the amount of code and the amount of distortion, and sends the merge index indicating the selected merge candidate and the inter prediction information of the merge candidate to the motion compensation prediction unit 306. supply. On the other hand, the decoding-side merge candidate selection unit 447 selects a merge candidate based on the decoded merge index, and supplies the selected merge candidate to the motion compensation prediction unit 406 .

The normal merge mode derivation unit 302 and the normal merge mode derivation unit 402 derive merge candidates in the parent block of a coding block when the size (the product of width and height) of a coding block is less than 32. Then, all child blocks use the merge candidates derived in the parent block. However, this is limited to cases where the size of the parent block is 32 or more and fits within the screen.

<Derivation of sub-block prediction motion vector mode>
Sub-block predictor motion vector mode derivation will be described.

FIG. 26 is a block diagram of sub-block predicted motion vector mode deriving section 303 in the encoding apparatus of this embodiment.

First, an affine inherited motion vector predictor candidate derivation unit 361 derives an affine inherited motion vector predictor candidate. The details of the derivation of the affine inheritance motion vector predictor candidate will be described later.

Subsequently, an affine-constructed motion vector predictor candidate deriving unit 362 derives an affine-constructed motion vector predictor candidate. The details of the derivation of the affine-constructed motion vector predictor candidate will be described later.

Subsequently, an affine identical motion vector predictor candidate derivation unit 363 derives an affine identical motion vector predictor candidate. The details of the derivation of the affine same motion vector predictor candidate will be described later.

The sub-block motion vector detection unit 366 detects a sub-block motion vector suitable for the sub-block prediction motion vector mode, and supplies the detected vector to the sub-block prediction motion vector candidate selection unit 367 and difference calculation unit 368 .

The sub-block motion vector predictor candidate selection unit 367 selects sub-block motion vector predictor candidates derived by the affine succession motion vector predictor candidate derivation unit 361, the affine construction motion vector predictor candidate derivation unit 362, and the affine identical motion vector predictor candidate derivation unit 363. A sub-block motion vector predictor candidate is selected based on the motion vector supplied from the sub-block motion vector detection unit 366, and information about the selected sub-block predictor motion vector candidate is sent to the inter prediction mode determination unit 305, It is supplied to the difference calculator 368 .

The difference calculation unit 368 subtracts the sub-block prediction motion vector selected by the sub-block prediction motion vector candidate selection unit 367 from the motion vector supplied from the sub-block motion vector detection unit 366, and calculates the difference prediction motion vector as an inter It is supplied to the prediction mode determination unit 305 .

FIG. 27 is a block diagram of sub-block predictive motion vector mode deriving section 403 in the decoding device of this embodiment.

First, an affine inherited motion vector predictor candidate derivation unit 461 derives an affine inherited motion vector predictor candidate. The processing of the affine inherited motion vector predictor candidate deriving unit 461 is the same as the processing of the affine inherited motion vector predictor candidate deriving unit 361 in the encoding apparatus of this embodiment.

Subsequently, an affine-constructed motion vector predictor candidate derivation unit 462 derives an affine-constructed motion vector predictor candidate. The processing of the affine-constructed motion vector predictor candidate derivation unit 462 is the same as the processing of the affine-constructed motion vector predictor candidate derivation unit 362 in the encoding apparatus of this embodiment.

Subsequently, the same affine motion vector predictor candidate derivation unit 463 derives the same affine motion vector predictor candidate. The processing of the affine identical motion vector predictor candidate deriving unit 463 is the same as the processing of the affine identical motion vector predictor candidate deriving unit 363 in the encoding apparatus of this embodiment.

The sub-block motion vector predictor candidate selection unit 466 selects sub-block motion vector predictor candidates derived by the affine succession motion vector predictor candidate derivation unit 461, the affine construction motion vector predictor candidate derivation unit 462, and the affine identical motion vector predictor candidate derivation unit 463. , a sub-block motion vector predictor candidate is selected based on the motion vector predictor index transmitted from the encoding device and decoded, and information on the selected sub-block motion vector predictor candidate is added to the motion compensation prediction unit 406. It is supplied to the calculation unit 467 .

The addition calculation unit 467 performs motion compensation prediction on a motion vector generated by adding a differential motion vector transmitted from the encoding device and decoded to the sub-block prediction motion vector selected by the sub-block prediction motion vector candidate selection unit 466. 406.

<Derivation of Affine Succession Predictor Motion Vector Candidates>
The affine succession motion vector predictor candidate derivation unit 361 will be described. The affine inherited motion vector predictor candidate deriving unit 461 is similar to the affine inherited motion vector predictor candidate deriving unit 361 .

The affine-inherited motion vector predictor candidate inherits the motion vector information of the control point.

FIG. 30 is a diagram illustrating derivation of affine inheritance motion vector predictor candidates.

Affine inherited motion vector predictor candidates are obtained by searching motion vectors of control points of spatially adjacent encoded/decoded blocks.

Specifically, a maximum of one affine mode is searched from each of the blocks (A0, A1) adjacent to the left side of the block to be processed and the blocks (B0, B1, B2) adjacent to the block to be processed above. , affine-inherited motion vector prediction.

FIG. 34 is a flowchart for deriving affine inherited motion vector predictor candidates.

First, the block (A0, A1) adjacent to the left side of the block to be processed is set as a left group (step S3101), and it is determined whether or not the block including A0 is a block using affine transform motion compensation (affine mode). (step S3102). If A0 is the affine mode (step S3102: YES), the affine mode used by A0 is obtained (step S3103), and the process moves to the upper adjacent block. If A0 is not the affine mode (step S3102: NO), the affine succession motion vector predictor candidate derivation target is A0->A1, and the acquisition of the affine mode is attempted from the block containing A1.

Subsequently, blocks (B0, B1, B2) adjacent to the upper side of the block to be processed are set as an upper group (step S3104), and it is determined whether or not the block including B0 is in the affine mode (step S3105). If B0 is the affine mode (step S3105: YES), the affine mode used by B0 is obtained (step S3106), and the process ends. If B0 is not the affine mode (step S3105: NO), the affine succession motion vector predictor candidate derivation target is B0->B1, and an attempt is made to acquire the affine mode from the block containing B1. Furthermore, if B1 is not the affine mode (step S3105: NO), the target for affine succession motion vector predictor candidate derivation is B1->B2, and an attempt is made to acquire the affine mode from the block containing B2.

In this way, the groups are divided into the left block and the upper block, and the left block is searched for affine modes in order from the lower left block to the upper left block, and the left block is searched for affine modes in order from the upper right block to the upper left block. As a result, two different affine modes can be obtained as much as possible, and an affine predictor motion vector candidate with a smaller differential motion vector can be derived.

<Derivation of Affine Construction Predictor Motion Vector Candidates>
The affine-constructed motion vector predictor candidate derivation unit 362 will be described. The affine-constructed motion vector predictor candidate derivation unit 462 is similar to the affine-constructed motion vector predictor candidate derivation unit 362 .

Affine-constructed motion vector predictor candidates construct motion vector information for control points from motion information for spatially adjacent blocks.

FIG. 31 is a diagram for explaining the derivation of affine-constructed motion vector predictor candidates.

An affine-constructed motion vector predictor candidate is obtained by constructing a new affine mode by combining motion vectors of spatially adjacent encoded/decoded blocks.

Specifically, the motion vector of the upper left control point CP0 is derived from the block (B2, B3, A2) adjacent to the upper left side of the block to be processed, and the block (B1, B0 ), and the motion vector of the lower left control point CP2 is derived from the block (A1, A0) adjacent to the lower left side of the block to be processed.

FIG. 35 is a flowchart for deriving affine-constructed motion vector predictor candidates.

First, the upper left control point CP0, the upper right control point CP1, and the lower left control point CP2 are derived (step S3201). The upper left control point CP0 is calculated by searching for reference blocks having the same reference image as the block to be processed in the order of priority of the B2, B3, and A2 reference blocks. The upper right control point CP1 is calculated by searching for a reference block having the same reference image as the block to be processed in the order of priority of the B1 and B0 reference blocks. The lower left control point CP2 is calculated by searching for a reference block having the same reference image as the block to be processed in the order of priority of the A1 and A0 reference blocks.

If the three control point mode is selected as the affine constructed predicted motion vector (step S3202: YES), whether or not all three control points (top left control point CP0, top right control point CP1, bottom left control point CP2) have been derived is determined (step S3203). When all three control points (top left control point CP0, top right control point CP1, bottom left control point CP2) are derived (step S3203: YES), three control points (top left control point CP0, top right control point CP1, bottom left control point CP1) The affine model using the point CP2) is set as the affine-constructed predicted motion vector (step S3204). If the 2 control point mode is selected instead of the 3 control point mode (step S3202: NO), it is determined whether or not all the 2 control points (upper left control point CP0, upper right control point CP1) have been derived. (step S3205). When all the two control points (top left control point CP0, top right control point CP1) are derived (step S3205: YES), the affine model using the two control points (top left control point CP0, top right control point CP1) is It is set as a constructed predictive motion vector (step S3206).

<Derivation of Affine Same Predictor Motion Vector Candidates>
The affine identical motion vector predictor candidate derivation unit 363 will be described. The affine identical motion vector predictor candidate derivation unit 463 is similar to the affine identical motion vector predictor candidate derivation unit 363 .

Affine same motion vector predictor candidates are obtained by deriving the same motion vector at each control point.

Specifically, similar to the affine-constructed motion vector predictor candidate deriving units 362 and 462, each control point information is derived, and all control points are set to be the same one of CP0 to CP2. It can also be obtained by setting temporal motion vectors derived in the same manner as in the normal motion vector predictor mode to all control points.

<Derivation of sub-block merge mode>
Sub-block merging mode derivation is described.

FIG. 28 is a block diagram of sub-block merge mode deriving section 304 in the encoding apparatus of this embodiment. The sub-block merge mode derivation unit 304 has a sub-block merge candidate list subblockMergeCandList. This has a list structure similar to the merge candidate list mergeCandList in the normal merge mode derivation unit 302, and stores the merge index indicating the location inside the sub-block merge candidate list and the sub-block merge candidate corresponding to the index as elements. A storage area is provided. In this embodiment, the subblock merge candidate list subblockMergeCandList can register at least five merge candidates (inter prediction information). However, each merging candidate further has motion vector information for each subblock or motion vector information for control points.

First, the sub-block temporal merge candidate derivation unit 381 derives sub-block temporal merge candidates. The details of sub-block temporal merging candidate derivation will be described later.

Subsequently, an affine succession merge candidate derivation unit 382 derives affine succession merge candidates. Details of affine inheritance merge candidate derivation will be described later.

Subsequently, an affine-construction merge candidate derivation unit 383 derives affine-construction merge candidates. The details of derivation of affine construction merge candidates will be described later.

Subsequently, the affine fixed merge candidate derivation unit 385 derives affine fixed merge candidates. Details of affine fixed merge candidate derivation will be described later.

The sub-block merge candidate selection unit 386 selects sub-block merge candidates derived by the sub-block temporal merge candidate derivation unit 381, the affine succession merge candidate derivation unit 382, the affine construction merge candidate derivation unit 383, and the affine fixed merge candidate derivation unit 385. A sub-block merging candidate is selected from among them, and information about the selected sub-block merging candidate is supplied to the inter-prediction mode determination unit 305 .

FIG. 29 is a block diagram of sub-block merging mode derivation section 404 in the decoding device of this embodiment. The subblock merge mode deriving unit 404 has a subblock merge candidate list subblockMergeCandList. This is the same as sub-block merging mode deriving section 304 .

First, the sub-block temporal merge candidate derivation unit 481 derives sub-block temporal merge candidates. The processing of the sub-block temporal merge candidate derivation section 481 is the same as the processing of the sub-block temporal merge candidate derivation section 381 .

Subsequently, an affine succession merge candidate derivation unit 482 derives affine succession merge candidates. The processing of the affine succession merge candidate derivation unit 482 is the same as the processing of the affine succession merge candidate derivation unit 382 .

Subsequently, an affine-construction merge candidate derivation unit 483 derives affine-construction merge candidates. The processing of the affine-construction merge candidate derivation unit 483 is the same as the processing of the affine-construction merge candidate derivation unit 383 .

Subsequently, an affine fixed merge candidate derivation unit 485 derives affine fixed merge candidates. The processing of the affine fixed merge candidate derivation unit 485 is the same as the processing of the affine fixed merge candidate derivation unit 485 .

The sub-block merge candidate selection unit 486 selects sub-block merge candidates derived by the sub-block temporal merge candidate derivation unit 481, the affine succession merge candidate derivation unit 482, the affine construction merge candidate derivation unit 483, and the affine fixed merge candidate derivation unit 485. A sub-block merging candidate is selected from among them based on the index transmitted and decoded from the encoder, and information about the selected sub-block merging candidate is provided to the motion compensated prediction unit 406 .

Sub-block merging mode deriving section 304 and sub-block merging mode deriving section 404 determine that, when the size of an encoding block (the product of width and height) is less than 32, sub-block merging candidates are found in the parent block of the encoding block. derived. Then, all child blocks use the sub-block merging candidates derived in the parent block. However, this is limited to cases where the size of the parent block is 32 or more and fits within the screen.

<Derivation of sub-block time merge candidates>
The operation of the sub-block temporal merge candidate derivation unit 381 will be described later.

<Affine Inheritance Merge Candidate Derivation>
The affine inheritance merge candidate deriving unit 382 will be described. The affine succession merge candidate derivation unit 482 is similar to the affine succession merge candidate derivation unit 382 .

Affine inheritance merge candidates inherit the affine models of control points from the affine models of spatially adjacent blocks.

FIG. 32 is a diagram for explaining derivation of affine inheritance merge candidates. Derivation of affine merge succession merge mode candidates is obtained by searching motion vectors of control points of spatially adjacent coded/decoded blocks, similar to derivation of affine inheritance prediction motion vectors.

Specifically, a maximum of one affine mode is searched from each of the blocks (A0, A1) adjacent to the left side of the block to be processed and the blocks (B0, B1, B2) adjacent to the block to be processed above. , used for affine merge mode.

FIG. 36 is a flowchart for deriving affine inheritance merge candidates.

First, the block (A0, A1) adjacent to the left side of the block to be processed is taken as the left group (step S3301), and it is determined whether or not the block including A0 is in the affine mode (step S3302). If A0 is in the affine mode (step S3102: YES), the affine model used by A0 is obtained (step S3303), and the process moves to the upper adjacent block. If A0 is not an affine mode (step S3302: NO), the target of affine inheritance merging candidate derivation is A0->A1, and an attempt is made to acquire an affine mode from a block containing A1.

Subsequently, blocks (B0, B1, B2) adjacent to the upper side of the block to be processed are set as an upper group (step S3304), and it is determined whether or not the block including B0 is in the affine mode (step S3305). If B0 is in the affine mode (step S3305: YES), the affine model used by B0 is obtained (step S3306), and the process ends. If B0 is not an affine mode (step S3305: NO), the target of affine inheritance merging candidate derivation is B0->B1, and an attempt is made to acquire an affine mode from a block including B1. Furthermore, if B1 is not an affine mode (step S3305: NO), the target of affine inheritance merging candidate derivation is B1->B2, and an attempt is made to obtain an affine mode from a block including B2.

<Derivation of affine construction merge candidates>
The affine construction merge candidate deriving unit 383 will be described. The affine-construction merge candidate derivation unit 483 is similar to the affine-construction merge candidate derivation unit 383 .

FIG. 33 is a diagram illustrating derivation of affine construction merge candidates. Affine building merge candidates build an affine model from temporally encoded blocks and motion information of spatially adjacent blocks.

Specifically, the motion vector of the upper left control point CP0 is derived from the block (B2, B3, A2) adjacent to the upper left side of the block to be processed, and the block (B1, B0 ), the motion vector of the lower left control point CP2 is derived from the block (A1, A0) adjacent to the lower left side of the block to be processed, and the motion vector of the lower left control point CP2 is derived from the lower right side of the block to be processed. A motion vector of the lower right control point CP3 is derived from the coding block (T0) adjacent to .

FIG. 37 is a flowchart for deriving affine construction merge candidates.

First, the upper left control point CP0, the upper right control point CP1, the lower left control point CP2, and the lower right control point CP3 are derived (step S3401). The upper left control point CP0 is calculated by searching for blocks having motion information in the priority order of B2, B3, and A2 blocks. The upper right control point CP1 is calculated by searching for blocks having motion information in the priority order of B1 and B0 blocks. The lower left control point CP2 is calculated by searching blocks having motion information in the priority order of A1 and A0 blocks. The lower right control point CP3 is calculated by searching the motion information of the time block.

Subsequently, it is determined whether or not an affine model with three control points can be constructed from the derived upper left control point CP0, upper right control point CP1, and lower left control point CP2 (step S3402). Step S3402: YES), the three control point affine model consisting of the upper left control point CP0, the upper right control point CP1, and the lower left control point CP2 is set as an affine merge candidate (step S3403).

Subsequently, it is determined whether or not an affine model with three control points can be constructed from the derived upper left control point CP0, upper right control point CP1, and lower right control point CP3 (step S3404). (Step S3404: YES), the three control point affine model consisting of the upper left control point CP0, the upper right control point CP1, and the lower right control point CP3 is set as an affine merge candidate (step S3405).

Subsequently, it is determined whether or not an affine model with three control points can be constructed from the derived upper left control point CP0, lower left control point CP2, and lower right control point CP3 (step S3406). (Step S3406: YES), the three control point affine model consisting of the upper left control point CP0, the lower left control point CP2, and the lower right control point CP3 is set as an affine merge candidate (step S3407).

Next, it is determined whether or not an affine model with three control points can be constructed from the derived upper right control point CP1, lower left control point CP2, and lower right control point CP3 (step S3408). (Step S3408: YES), the three control point affine model consisting of the upper right control point CP1, the lower left control point CP2, and the lower right control point CP3 is set as an affine merge candidate (step S3409).

Subsequently, it is determined whether or not an affine model with two control points can be constructed from the derived upper left control point CP0 and upper right control point CP1 (step S3410), and if it can be constructed (step S3410: YES) , the upper left control point CP0 and the upper right control point CP1 are set as affine merge candidates (step S3411).

Subsequently, it is determined whether or not an affine model with two control points can be constructed from the derived upper left control point CP0 and lower left control point CP2 (step S3412), and if it can be constructed (step S3412: YES) , the upper left control point CP0 and the lower left control point CP2 are set as affine merge candidates (step S3413).

Here, whether or not an affine model can be constructed is at least conditional on the fact that the reference images of all control points are the same (affine transformation is possible). Affine models other than the three control point affine model with the upper left control point CP0, the upper right control point CP1 and the lower left control point CP2, and the two control point affine models with the upper left control point CP0 and the upper right control point CP1 are the three control affine models. The model is a three control point affine model with upper left control point CP0, upper right control point CP1 and lower left control point CP2, and the two control affine model is a two control point affine model with upper left control point CP0 and upper right control point CP1. Convert to model.

<Derivation of affine fixed merge candidates>
The affine fixed merge candidate derivation unit 385 will be described. The affine fixed merge candidate derivation unit 485 is similar to the affine fixed merge candidate derivation unit 385 .

Affine fixed merge candidates fix the motion information of control points with fixed motion information.

Specifically, the motion vector of each control point is fixed at (0, 0).

<Derivation of temporal motion vector prediction>
Prior to the description of the temporal motion vector predictor, the temporal relationship of pictures will be described with reference to FIG. FIG. 49(a) shows the relationship between an encoding block to be processed and an encoded picture temporally different from the picture to be processed. A specific coded picture referred to in the coding of the picture to be processed is defined as ColPic. ColPic is specified by syntax.

Also, FIG. 49(b) shows coded blocks existing in the same position as the target coded block and its vicinity in ColPic. However, the T0 and T1 encoded blocks shown in FIG. 49(b) are schematic, and the actual positions and sizes are not limited to this. Assume that the position of the encoding block to be processed is (xCb, yCb), the width is cbWidth, and the height is cbHeight. and,
xColBr = xCb + cbWidth
yColBr = yCb + cbHeight
Calculate The coded block on ColPic containing position ((xColBr >> 3) << 3, (yColBr >> 3) << 3) is T0. again,
xColCtr = xCb + (cbWidth >> 1)
yColCtr = yCb + (cbHeight >> 1)
Calculate The coded block on ColPic containing location ((xColCtr >> 3) << 3, (yColCtr >> 3) << 3) is T1.

Although the above description of the temporal context of pictures is for encoding, the same is true for decoding. In other words, the decoding is explained in the same way by replacing the encoding in the above explanation with the decoding.

The operation of the temporal motion vector predictor candidate derivation unit 322 in the normal motion vector predictor mode derivation unit 301 of FIG. 17 will be described with reference to FIG.

First, ColPic is derived (step S4201). Derivation of ColPic will be described with reference to FIG.

If the slice type slice_type is a B slice and the flag collocated_from_l0_flag is 0 (step S4211: YES, step S4212: YES), the picture ColPic at a different time is the picture RefPicList1[0] whose reference index is 0 in the reference list L1 ( step S4213). Otherwise, that is, if the slice type slice_type is B slice and the flag collocated_from_l0_flag is 1 (step S4211: YES, step S4212: NO), or if the slice type slice_type is P slice (step S4211: NO, step S4214: YES), the picture ColPic at a different time becomes the picture RefPicList0[0] whose reference index is 0 in the reference list L0 (step S4215). If the slice_type is not P slice (step S4214: NO), the process ends.

Again, refer to FIG. After deriving ColPic, the coded block colCb is derived to obtain the coded information (step S4202). This processing will be described with reference to FIG.

First, in a picture ColPic at a different time, a coded block including the lower right position at the same position as the coding block to be coded is set as a coded block colCb at a different time (step S4221). An example of this encoding block is shown in encoding block T0 in FIG.

Next, the coded information of the coded block colCb at a different time is obtained (step S4222). If the PredMode of the coded block colCb at a different time cannot be used or the prediction mode PredMode of the coded block colCb at a different time is intra prediction (MODE_INTRA) (step S4223: NO, step S4224: YES), the picture at the different time A coded block colCb at a different time is defined as a coded block colCb at the same position as the target coded block in ColPic (step S4225). An example of this encoding block is shown in encoding block T1 in FIG.

Again, refer to FIG. Next, inter prediction information is derived for each reference list (steps S4203 and S4204). Here, for the coding block colCb, the motion vector mvLXCol for each reference list and the flag availableFlagLXCol indicating whether or not the coding information is valid are derived. LX indicates a reference list, and in the derivation of reference list 0, LX becomes L0, and in the derivation of reference list 1, LX becomes L1. Derivation of inter prediction information will be described with reference to FIG.

If the coded block colCb at a different time cannot be used (step S4231: NO), or if the prediction mode PredMode is intra prediction (MODE_INTRA) (step S4232: NO), both flag availableFlagLXCol and flag predFlagLXCol are set to 0 (step S4233). , the motion vector mvLXCol is set to (0, 0) (step S4234), and the process ends.

If the coding block colCb is available (step S4231: YES) and the prediction mode PredMode is not intra prediction (MODE_INTRA) (step S4232: YES), mvCol, refIdxCol and availableFlagCol are calculated in the following procedure.

If the flag PredFlagL0[xPCol][yPCol] indicating whether or not the L0 prediction of the coding block colCb is used is 0 (step S4235: YES), the prediction mode of the coding block colCb is Pred_L1. mvCol is set to the same value as the L1 motion vector MvL1[xPCol][yPCol] of the coding block colCb (step S4236), and the reference index refIdxCol is set to the same value as the L1 reference index RefIdxL1[xPCol][yPCol]. is set (step S4237), and the reference list listCol is set to L1 (step S4238). where xPCol, yPCol are indices indicating the position of the upper left pixel of the coded block colCb within the picture ColPic at different times.

On the other hand, if the L0 prediction flag PredFlagL0[xPCol][yPCol] of the coding block colCb is not 0 (step S4235: NO), it is determined whether or not the L1 prediction flag PredFlagL1[xPCol][yPCol] of the coding block colCb is 0. do. When the L1 prediction flag PredFlagL1[xPCol][yPCol] of the coding block colCb is 0 (step S4239: YES), the motion vector mvCol is the same as the L0 motion vector MvL0[xPCol][yPCol] of the coding block colCb. value (step S4240), the reference index refIdxCol is set to the same value as the reference index RefIdxL0[xPCol][yPCol] of L0 (step S4241), and the reference list listCol is set to L0 (step S4242).

If both the L0 prediction flag PredFlagL0[xPCol][yPCol] of the encoding block colCb and the L1 prediction flag PredFlagL1[xPCol][yPCol] of the encoding block colCb are not 0 (step S4235: NO and S4239: NO), encoding Since the inter prediction mode of block colCb is bi-prediction (Pred_BI), one of the two motion vectors L0 and L1 is selected (step S4243).

FIG. 54 is a flowchart showing a derivation processing procedure of inter prediction information of a coding block when the inter prediction mode of the coding block colCb is bi-prediction (Pred_BI).

First, it is determined whether or not the POCs of all pictures registered in all reference lists are smaller than the POC of the current picture to be processed (step S4251). When the POCs of all pictures registered in L1 are smaller than the POC of the current picture to be processed (step S4251: YES), LX is L0, that is, the motion vector predictor candidate of the L0 motion vector of the coding block to be processed. is derived (step S4252: YES), the L0 inter prediction information of the coding block colCb is selected, and LX is L1, that is, the prediction vector candidate of the L1 motion vector of the coding block to be processed is If so (step S4252: NO), the L1 inter prediction information of the coding block colCb is selected. On the other hand, if at least one of the POCs of the pictures registered in the reference lists L0 and L1 of the coding block colCb is greater than the POC of the current picture to be processed (step S4251: NO), and the flag collocated_from_l0_flag is 0 ( If the flag collocated_from_l0_flag is 1 (step S4253: NO), the L1 inter prediction information of the coding block colCb is selected. .

If the L0 inter prediction information of the coding block colCb is selected (step S4252: YES or step S4253: YES), the motion vector mvCol is set to the same value as MvL0[xPCol][yPCol] (step S4254). , the reference index refIdxCol is set to the same value as RefIdxL0[xPCol][yPCol] (step S4255), and the list listCol is set to L0 (step S4256).

When selecting the L1 inter prediction information of the coding block colCb (step S4252: NO or step S4253: NO), the motion vector mvCol is set to the same value as MvL1[xPCol][yPCol] (step S4257). , the reference index refIdxCol is set to the same value as RefIdxL1[xPCol][yPCol] (step S4258), and the list listCol is set to L1 (step S4259).

Returning to FIG. 53, when the inter prediction information is obtained from the coding block colCb, both the flag availableFlagLXCol and the flag predFlagLXCol are set to 1 (step S4244).

Subsequently, the motion vector mvCol is scaled to obtain a motion vector mvLXCol (step S4245). The scaling operation processing procedure for this motion vector mvLXCol will be described with reference to FIG.

The inter-picture distance td is obtained by subtracting the POC of the reference picture corresponding to the reference index refIdxCol referenced in the list listCol of the coding block colCb from the POC of the picture ColPic at a different time,
td = [POC of picture ColPic at different time]-[POC of reference picture referenced in list listCol of coding block colCb]
is calculated (step S4261). Note that when the POC of the reference picture referenced by the list listCol of the coding block colCb is earlier than the picture ColPic at a different time in the display order, the inter-picture distance td is a positive value, and the picture ColPic at a different time is earlier than the picture ColPic at a different time. If the POC of the reference picture referred to by the list listCol of the coding block colCb is later in the display order, the inter-picture distance td is a negative value.

Next, the POC of the reference picture referred to by the list LX of the current picture to be processed is subtracted from the POC of the current picture to be processed to obtain the inter-picture distance tb,
tb = [POC of current picture being processed] - [POC of reference picture corresponding to reference index of LX of temporal merge candidate]
is calculated (step S4262). Note that when the reference picture referred to in the list LX of the current picture to be processed is earlier than the current picture to be processed in the display order, the inter-picture distance tb becomes a positive value and the list LX of the current picture to be processed. If the reference picture referred to by is later in the display order, the inter-picture distance tb is a negative value.

Subsequently, the inter-picture distances td and tb are compared (step S4263), and if the inter-picture distances td and tb are equal (step S4263: YES), the motion vector mvLXCol is
mvLXCol = mvCol
is calculated (step S4264), and the scaling calculation process ends.

On the other hand, if the inter-picture distance td and tb are not equal (step S4263: NO), the variable tx is
tx = ( 16384 + Abs( td ) >> 1 ) / td
is calculated (step S4265). Then, set the scaling factor distScaleFactor to
distScaleFactor = Clip3( -4096, 4095, ( tb * tx + 32 ) >> 6 )
is calculated (step S4266). Here, Clip3(x,y,z) is a function that limits the minimum value to x and the maximum value to y for the value z. Subsequently, the motion vector mvLXCol is
mvLXCol = Clip3( -32768, 32767, Sign( distScaleFactor * mvLXCol )
* ( (Abs( distScaleFactor * mvLXCol ) + 127 ) >> 8 ) )
is calculated (step S4267), and the scaling calculation process ends. Here, Sign(x) is a function that returns the sign of value x, and Abs(x) is a function that returns the absolute value of value x.

Again, refer to FIG. Then, the L0 motion vector mvL0Col is added as a candidate to the motion vector predictor candidate list mvpListLX in the normal motion vector predictor mode deriving unit 301 (step S4205). However, this addition is only when the flag availableFlagL0Col=1 indicating whether or not the coded block colCb in the reference list 0 is valid. Also, the L1 motion vector mvL1Col is added as a candidate to the motion vector predictor candidate list mvpListLX in the normal motion vector predictor mode deriving unit 301 (step S4205). However, this addition is only when the flag availableFlagL1Col=1 indicating whether or not the coded block colCb in reference list 1 is valid. Thus, the processing of the temporal motion vector predictor candidate deriving unit 322 ends.

The above description of the normal motion vector predictor mode derivation unit 301 is for encoding, but the same applies to decoding. In other words, the operation of the temporal motion vector predictor candidate derivation unit 422 in the normal motion vector predictor mode derivation unit 401 in FIG. 23 will be described in the same way, substituting decoding for encoding in the above description.

<Derivation of time merge candidates>
The operation of the temporal merge candidate derivation unit 342 in the normal merge mode derivation unit 302 of FIG. 18 will be described with reference to FIG.

First, ColPic is derived (step S4301). Next, a coded block colCb is derived and coded information is obtained (step S4302). Furthermore, inter prediction information is derived for each reference list (steps S4303 and S4304). The above processing is the same as S4201 to S4204 in the temporal motion vector predictor candidate derivation unit 322, and therefore description thereof is omitted.

Next, a flag availableFlagCol indicating whether or not the encoded block colCb is valid is calculated (step S4305). availableFlagCol is 1 when flag availableFlagL0Col or flag availableFlagL1Col is 1. otherwise availableFlagCol is 0.

Then, the L0 motion vector mvL0Col and the L1 motion vector mvL1Col are added as candidates to the merge candidate list mergeCandList in the normal merge mode derivation unit 302 (step S4306). However, this addition is only when the flag availableFlagCol=1, which indicates whether or not the encoded block colCb is valid. Thus, the processing of the temporal merge candidate derivation unit 342 is completed.

The above description of the temporal merge candidate deriving unit 342 is for encoding, but the same applies to decoding. That is, the operation of the temporal merge candidate derivation unit 442 in the normal merge mode derivation unit 402 of FIG. 24 will be similarly described by replacing the encoding with decoding in the above description.

<Update history motion vector predictor candidate list>
Next, a method for initializing and updating the history motion vector predictor candidate list HmvpCandList provided in the encoding information storage memory 111 on the encoding side and the encoding information storage memory 205 on the decoding side will be described in detail. FIG. 38 is a flowchart for explaining the history motion vector predictor candidate list initialization/update processing procedure.

In this embodiment, it is assumed that the history motion vector predictor candidate list HmvpCandList is updated by the encoding information storage memory 111 and the encoding information storage memory 205 . A history candidate list update unit may be installed in the inter prediction unit 102 and the inter prediction unit 203 to update the history motion vector predictor candidate list HmvpCandList.

The historical motion vector predictor candidate list HmvpCandList is initialized at the beginning of the slice, and the historical motion vector predictor candidate list HmvpCandList is updated when the prediction method determination unit 105 selects the normal vector predictor mode or the normal merge mode on the encoding side. On the decoding side, the history motion vector predictor candidate list HmvpCandList is updated when the inter prediction mode decoded by the bitstream decoding unit 201 is the normal vector predictor mode or the normal merge mode.

Inter prediction information used when inter prediction is performed in the normal vector predictor mode or normal merge mode is registered in the history motion vector predictor candidate list HmvpCandList as the inter prediction information candidate hMvpCand. Inter prediction information candidate hMvpCand includes L0 reference index refIdxL0 and L1 reference index refIdxL1, L0 prediction flag predFlagL0 indicating whether L0 prediction is performed, L1 prediction flag predFlagL1 indicating whether L1 prediction is performed, A motion vector mvL0 of L0 and a motion vector mvL1 of L1 are included. Inter prediction information candidate If inter prediction information with the same value as hMvpCand exists, that element is deleted from the historical motion vector predictor candidate list HmvpCandList. On the other hand, if there is no inter prediction information with the same value as the inter prediction information candidate hMvpCand, the leading element of the historical motion vector predictor candidate list HmvpCandList is deleted, and the inter prediction information candidate Add hMvpCand.

The number of elements of the history motion vector predictor candidate list HmvpCandList provided in the encoding information storage memory 111 on the encoding side and the encoding information storage memory 205 on the decoding side of the present invention is six.

First, the history motion vector predictor candidate list HmvpCandList is initialized for each slice (step S2101 in FIG. 38). At the beginning of the slice, all elements of the historical motion vector predictor candidate list HmvpCandList are emptied, and the value of NumHmvpCand, the number of historical motion vector predictor candidates registered in the historical motion vector predictor candidate list HmvpCandList, is set to zero.

Although the historical motion vector predictor candidate list HmvpCandList is initialized in units of slices (first coded blocks in slices), it may be initialized in units of pictures, tiles, or treeblock rows.

Subsequently, the following updating process of the history motion vector predictor candidate list HmvpCandList is repeatedly performed for each encoded block in the slice (steps S2102 to S2107 in FIG. 38).

First, initialization is performed for each encoding block. A flag identicalCandExist indicating whether or not the same candidate exists is set to FALSE, and a deletion target index removeIdx is set to 0 (step S2103 in FIG. 38).

It is determined whether or not the inter prediction information candidate hMvpCand to be registered exists in the historical motion vector predictor candidate list HmvpCandList (step S2104 in FIG. 38). When the prediction method determination unit 105 on the encoding side determines the normal motion vector prediction mode or the normal merge mode, or when the bit stream decoding unit 201 on the decoding side decodes as the normal motion vector prediction mode or the normal merge mode, Let the inter prediction mode be hMvpCand. When the prediction method determination unit 105 on the encoding side determines the intra prediction mode, the sub-block motion vector prediction mode, or the sub-block merge mode, or when the bitstream decoding unit 201 on the decoding side selects the intra prediction mode or the sub-block prediction motion vector mode. Alternatively, when decoding is performed in sub-block merge mode, the history motion vector predictor candidate list HmvpCandList is not updated, and there is no inter prediction information candidate hMvpCand to be registered. If there is no inter prediction information candidate hMvpCand to be registered, steps S2105 and S2106 are skipped (step S2104 in FIG. 38: NO). If there is an inter-prediction information candidate hMvpCand to be registered, the process from step S2105 is performed (step S2104 in FIG. 38: YES).

Subsequently, it is determined whether or not each element of the historical motion vector predictor candidate list HmvpCandList includes the same element as the inter prediction information candidate hMvpCand to be registered (step S2105 in FIG. 38). FIG. 39 is a flow chart of this identical element confirmation processing procedure. If the value of the number of historical motion vector predictor candidates NumHmvpCand is 0 (step S2121 in FIG. 39: NO), the historical motion vector predictor candidate list HmvpCandList is empty and the same candidate does not exist, so steps S2122 to S2125 in FIG. 39 are skipped. , ends the same element confirmation processing procedure. If the value of the number of historical motion vector predictor candidates NumHmvpCand is greater than 0 (step S2121 in FIG. 39: YES), the processing in step S2123 is repeated until the historical motion vector predictor index hMvpIdx is from 0 to NumHmvpCand-1 (step S2123 in FIG. 39). S2122-S2125). First, it is compared whether or not the hMvpIdx-th element HmvpCandList[hMvpIdx] counting from 0 in the historical motion vector predictor candidate list is the same as the inter prediction information candidate hMvpCand (step S2123 in FIG. 39). If they are the same (step S2123 in FIG. 39: YES), the flag identicalCandExist indicating whether or not the same candidate exists is set to TRUE, the deletion target index removeIdx is set to the hMVpIndex value, and the same candidate is set. Terminate the element confirmation process. If they are not the same (step S2123 in FIG. 39: NO), hMvpIdx is incremented by 1, and if the historical motion vector index hMvpIdx is less than or equal to NumHmvpCand-1, the processes after step S2123 are performed (steps S2122 to S2125 in FIG. 39). .

Returning to the flowchart of FIG. 38 again, shift and addition processing of the elements of the historical motion vector predictor candidate list HmvpCandList is performed (step S2106 of FIG. 38). FIG. 40 is a flowchart of the element shift/addition processing procedure of the historical motion vector predictor candidate list HmvpCandList in step S2106 of FIG. First, it is determined whether to add a new element after removing the elements stored in the historical motion vector predictor candidate list HmvpCandList or to add a new element without removing the elements. Specifically, it is compared whether the flag identicalCandExist indicating whether or not the same candidate exists is TRUE (true) or whether NumHmvpCand is 6 (step S2141 in FIG. 40). If the flag identicalCandExist indicating whether or not the same candidate exists is TRUE (true) or if NumHmvpCand satisfies either condition of 6 (step S2141 in FIG. 40: YES), the motion vector predictor candidate list HmvpCandList stores Remove existing elements and add new elements. Set the initial value of index i to the value of removeIdx + 1. The element shift processing in step S2143 is repeated from this initial value to NumHmvpCand. (Steps S2142 to S2144 in FIG. 40). By copying the element of HMVPCandList[i] to HMVPCandList[i-1], the element is shifted forward (step S2143 in FIG. 40), and i is incremented by 1 (steps S2142 to S2144 in FIG. 40). When the index i becomes NumHmvpCand+1 and the element shift processing in step S2143 is completed, the inter prediction information candidate hMvpCand is added to the end of the history motion vector predictor candidate list (step S2145 in FIG. 40). Here, the end of the historical motion vector predictor candidate list is the (NumHmvpCand-1)-th HMVPCandList[NumHmvpCand-1] counted from 0. This completes the element shift/addition process for the historical motion vector predictor candidate list HMVPCandList. On the other hand, if the flag identicalCandExist indicating whether or not the same candidate exists is TRUE (true) and NumHmvpCand does not satisfy the condition of 6 (step S2141 in FIG. 40: NO), it is stored in the historical motion vector predictor candidate list HmvpCandList. The inter-prediction information candidate hMvpCand is added to the end of the history motion vector predictor candidate list without removing the elements that are indicated (step S2146 in FIG. 40). Here, the end of the historical motion vector predictor candidate list is the NumHmvpCand-th HMVPCandList[NumHmvpCand] counted from 0. Also, NumHmvpCand is incremented by 1, and the element shift/addition processing of this historical motion vector predictor candidate list HMVPCandList is completed.

FIG. 43 is a diagram illustrating an example of update processing of the historical motion vector predictor list. When adding new inter prediction information when six elements (inter prediction information) are registered in the historical motion vector predictor candidate list HMVPCandList, each element of the historical motion vector predictor candidate list HMVPCandList and the new inter prediction information are added from the front. After comparing the prediction information (FIG. 43(a)), if the new inter prediction information has the same value as the third element HMVP2 from the top of the historical motion vector predictor candidate list HMVPCandList, Delete the element HMVP2, shift (copy) the backward elements HMVP3 to HMVP5 forward one by one, and add new inter prediction information to the end of the historical motion vector predictor candidate list HMVPCandList (Fig. 43(b)). , completes the update of the historical motion vector predictor candidate list HMVPCandList (FIG. 43(c)).

<Historical Motion Vector Predictor Candidate Derivation Processing>
Next, the historical motion vector predictor candidate deriving unit 323 of the normal predictive motion vector mode deriving unit 301 on the encoding side and the historical motion vector predictor candidate deriving unit 423 of the normal predictive motion vector mode deriving unit 401 on the decoding side perform common processing. A method for deriving historical motion vector predictor candidates from the historical motion vector predictor candidate list HMVPCandList, which is the processing procedure of step S304 in FIG. 20, will be described in detail. FIG. 41 is a flowchart for explaining the historical motion vector predictor candidate derivation processing procedure.

If the number of current motion vector predictor candidates numCurrMvpCand is greater than or equal to the maximum number of elements (here, assumed to be 2) in the motion vector predictor candidate list mvpListLX or the value of the number of historical motion vector predictor candidates NumHmvpCand is 0 (step S2201 in FIG. 41: NO), the processing of steps S2202 to S2209 in FIG. 41 is omitted, and the history motion vector predictor candidate derivation processing procedure ends. If the number of current motion vector predictor candidates numCurrMvpCand is less than 2, which is the maximum number of elements in the motion vector predictor candidate list mvpListLX, and if the value of the number of historical motion vector predictor candidates NumHmvpCand is greater than 0 (step S2201 in FIG. 41: YES), the processing of steps S2202 to S2209 in FIG. 41 is performed.

Subsequently, the processing from steps S2203 to S2208 in FIG. 41 is repeated until the index i is from 0 to 3 or the number of historical motion vector predictor candidates NumHmvpCand, whichever is smaller (steps S2202 to S2209 in FIG. 41). If the number of current motion vector predictor candidates numCurrMvpCand is equal to or greater than 2, which is the maximum number of elements in the motion vector predictor candidate list mvpListLX (step S2203 in FIG. 41: NO), the processing from steps S2204 to S2209 in FIG. The historical motion vector predictor candidate derivation processing procedure is ended. If the number of current motion vector predictor candidates numCurrMvpCand is less than 2, which is the maximum number of elements in the motion vector predictor candidate list mvpListLX (step S2203 in FIG. 41: YES), the processing from step S2204 onward in FIG. 41 is performed.

Subsequently, steps S2205 to S2207 are performed for Y=0 and 1 (L0 and L1) respectively (steps S2204 to S2208 in FIG. 41). If the number of current motion vector predictor candidates numCurrMvpCand is equal to or greater than 2, which is the maximum number of elements in the motion vector predictor candidate list mvpListLX (step S2205 in FIG. 41: NO), the processing from steps S2206 to S2209 in FIG. The historical motion vector predictor candidate derivation processing procedure is terminated. If the number of current motion vector predictor candidates numCurrMvpCand is less than 2, which is the maximum number of elements in the motion vector predictor candidate list mvpListLX (step S2205 in FIG. 41: YES), the processing from step S2206 onward in FIG. 41 is performed.

Subsequently, when the reference index of LY in the historical motion vector predictor candidate list HmvpCandList[i] is the same as the reference index refIdxLX of the motion vector to be encoded/decoded (step S2206 in FIG. 41: YES), the motion vector predictor candidate list As the last element, the LY motion vector of the historical motion vector predictor candidate HmvpCandList[i] is added to the numCurrMvpCand-th element mvpListLX[numCurrMvpCand] counted from 0 of the motion vector predictor candidate list (step S2207 in FIG. 41). numCurrMvpCand, which is the number of motion vector predictor candidates, is incremented by one. If the reference index of LY in the historical motion vector predictor candidate list HmvpCandList[i] is not the same as the reference index refIdxLX of the motion vector to be encoded/decoded (step S2206 in FIG. 41: NO), skip the additional processing in step S2207. .

The processes from steps S2205 to S2207 in FIG. 41 are performed for both L0 and L1 (steps S2204 to S2208 in FIG. 41).

The index i is incremented by 1, and if the index i is equal to or less than the smaller value of 3 or the number of historical motion vector predictor candidates NumHmvpCand, the processing from step S2203 onward is performed again (steps S2202 to S2209 in FIG. 41).

<History Merge Candidate Derivation Processing>
Next, the process of step S404 in FIG. 21, which is common to the history merge candidate derivation unit 345 of the normal merge mode derivation unit 302 on the encoding side and the history merge candidate derivation unit 445 of the normal merge mode derivation unit 402 on the decoding side. A method for deriving history merge candidates from the history merge candidate list HmvpCandList, which is a procedure, will be described in detail. FIG. 42 is a flowchart for explaining the history merging candidate derivation processing procedure.

First, initialization processing is performed (step S2301 in FIG. 42). Sets each element from 0 to (numCurrMergeCand -1) of isPruned[i] to the value FALSE, and sets the variable numOrigMergeCand to the number of elements registered in the current merge candidate list, numCurrMergeCand.

Subsequently, the initial value of the index hMvpIdx is set to 1, and the additional processing from steps S2303 to S2310 in FIG. 42 is repeated from this initial value to NumHmvpCand (steps S2302 to S2311 in FIG. 42). If the number of elements registered in the current merge candidate list, numCurrMergeCand, is not less than (the maximum number of merge candidates, MaxNumMergeCand-1), merge candidates have been added to all the elements in the merge candidate list. The process ends (step S2303 in FIG. 42: NO). If the number numCurrMergeCand of elements registered in the current merge candidate list is less than (maximum number of merge candidates MaxNumMergeCand-1) (step S2303 in FIG. 42: YES), the process from step S2304 onwards is performed.

First, sameMotion is set to FALSE (step S2304 in FIG. 42). Subsequently, the initial value of index i is set to 0, and steps S2306 and S2307 in FIG. 42 are performed from this initial value to 1 (S2305 to S2308 in FIG. 42).

Next, whether the (NumHmvpCand-hMvpIdx)-th element HmvpCandList[NumHmvpCand-hMvpIdx] of the history motion vector prediction candidate list and the i-th element mergeCandList[i] of the merge candidate list are the same value. It compares whether or not (step S2306 in FIG. 42). Here, the same value of the merge candidates means that the values of all components (inter prediction mode, reference index, motion vector) of the merge candidates are the same. However, this processing of step S2306 is limited to when hMvpIdx is greater than NumHmvpCand-2, mergeCandList[i] is a spatial merge candidate, and isPruned[i] is FALSE (false). If the values are the same (step S2306 in FIG. 39: YES), both sameMotion and isPruned[i] are set to TRUE (step S2307 in FIG. 42). If the values are not the same (step S2306 in FIG. 39: NO), the process of step S2307 is skipped. 42 is completed, it is compared whether sameMotion is FALSE (step S2309 in FIG. 42), and if sameMotion is FALSE (step S2309 in FIG. 42: YES), add the (NumHmvpCand - hMvpIdx)-th element HmvpCandList[NumHmvpCand - hMvpIdx] of the history motion vector predictor candidate list counting from 0 to the numCurrMergeCand-th mergeCandList[numCurrMergeCand] of the merge candidate list, and increment numCurrMergeCand by 1 ( step S2310 in FIG. 42). The index hMvpIdx is incremented by 1 (step S2302 in FIG. 42), and steps S2302 to S2311 in FIG. 42 are repeated.

When all elements of the history motion vector predictor candidate list have been confirmed or merge candidates have been added to all elements of the merge candidate list, this history merge candidate derivation process is completed.

<Average merging candidate derivation process>
Next, the processing of step S403 in FIG. 21, which is the processing common to the average merge candidate derivation unit 344 of the normal merge mode derivation unit 302 on the encoding side and the average merge candidate derivation unit 444 of the normal merge mode derivation unit 402 on the decoding side. The procedure for deriving average merge candidates will be described in detail. FIG. 62 is a flowchart for explaining the average merging candidate derivation processing procedure.

First, initialization processing is performed (step S1301 in FIG. 62). Set the number of elements registered in the current merge candidate list, numCurrMergeCand, to the variable numOrigMergeCand.

Subsequently, the merge candidate list is sequentially scanned from the top to determine two pieces of motion information. Let index i=0 indicating the first motion information and index j=1 indicating the second motion information. (Steps S1302 and S1303 in FIG. 62). If the number of elements registered in the current merge candidate list, numCurrMergeCand, is not less than (the maximum number of merge candidates, MaxNumMergeCand-1), merge candidates have been added to all the elements in the merge candidate list. The process ends (step S1304 in FIG. 62). If the number of elements registered in the current merge candidate list, numCurrMergeCand, is less than (the maximum number of merge candidates, MaxNumMergeCand-1), the process from step S1305 is performed.

It is determined whether or not both the i-th motion information mergeCandList[i] in the merge candidate list and the j-th motion information mergeCandList[j] in the merge candidate list are invalid (step S1305 in FIG. 62). If so, do not derive the average merge candidate for mergeCandList[i] and mergeCandList[j] and move on to the next element. If both mergeCandList[i] and mergeCandList[j] are not invalid, the following processing is repeated with X set to 0 and 1 (steps S1306 to S1314 in FIG. 62).

It is determined whether the LX prediction of mergeCandList[i] is valid (step S1307 in FIG. 62). If the LX prediction for mergeCandList[i] is valid, it is determined whether the LX prediction for mergeCandList[j] is valid (step S1308 in FIG. 62). If the LX prediction for mergeCandList[j] is valid, i.e. if both the LX prediction for mergeCandList[i] and the LX prediction for mergeCandList[j] are valid, then the motion vector for the LX prediction for mergeCandList[i] and mergeCandList Derive the average merge candidate for the LX prediction with the motion vector for the LX prediction that averaged the motion vector for the LX prediction in [j] and the reference index for the LX prediction in mergeCandList[i] and set it to the LX prediction for averageCand, LX prediction is validated (step S1309 in FIG. 62). In step S1308 of FIG. 62, if the LX prediction of mergeCandList[j] is not valid, that is, if the LX prediction of mergeCandList[i] is valid and the LX prediction of mergeCandList[j] is invalid, mergeCandList[i] LX prediction average merge candidate having a motion vector of LX prediction and reference index is derived, set to LX prediction of averageCand, and LX prediction of averageCand is validated (step S1310 in FIG. 62). If the LX prediction of mergeCandList[i] is not valid in step S1307 of FIG. 62, it is determined whether or not the LX prediction of mergeCandList[j] is valid (step S1311 of FIG. 62). If the LX prediction of mergeCandList[j] is valid, i.e. the LX prediction of mergeCandList[i] is invalid and the LX prediction of mergeCandList[j] is valid, then the motion vector of the LX prediction of mergeCandList[j] and An average merging candidate for LX predictions with a reference index is derived and set to averageCand's LX prediction, and averageCand's LX prediction is validated (step S1312 in FIG. 62). In step S1311 of FIG. 62, if the LX prediction for mergeCandList[j] is invalid, that is, if both the LX prediction for mergeCandList[i] and mergeCandList[j] are invalid, the LX prediction for averageCand is invalidated. (Step S1312 in FIG. 62).

The average merge candidate averageCand of the L0 prediction, L1 prediction or BI prediction generated as described above is added to the numCurrMergeCand-th mergeCandList[numCurrMergeCand] of the merge candidate list, and numCurrMergeCand is incremented by 1 (step S1315 in FIG. 62). This completes the average merging candidate derivation process.

Note that the average merge candidate is averaged for each of the horizontal component of the motion vector and the vertical component of the motion vector.

<Derivation of sub-block time merge candidates>
The operation of sub-block temporal merge candidate derivation section 381 in sub-block merge mode derivation section 304 of FIG. 16 will be described with reference to FIG.

First, it is determined whether or not the encoded block is less than 8×8 pixels (step S4002).

If the encoded block is less than 8×8 pixels (step S4002: YES), the flag availableFlagSbCol=0 indicating the existence of the sub-block temporal merge candidate is set (step S4003), and the processing of the sub-block temporal merge candidate derivation unit is terminated. . Here, if temporal motion vector prediction is prohibited by the syntax, or if sub-block time merging is prohibited, the same processing as when the coding block is less than 8×8 pixels (step S4002: YES) is performed. do.

On the other hand, if the encoded block is 8×8 pixels or more (step S4002: NO), adjacent motion information of the encoded block in the encoded picture is derived (step S4004).

Processing for deriving adjacent motion information of a coding block will be described with reference to FIG. The process of deriving adjacent motion information is similar to the process of the spatial motion vector predictor candidate derivation unit 321 described above. However, the order in which adjacent blocks are searched is A0, B0, B1, and A1, and B2 is not searched. First, encoding information is obtained with the adjacent block n=A0 (step S4052). The encoding information indicates a flag availableFlagN indicating whether or not adjacent blocks can be used, a reference index refIdxLXN for each reference list, and a motion vector mvLXN.

Next, it is determined whether the adjacent block n is valid or invalid (step S4054). If the flag availableFlagN=1 indicating whether or not the adjacent block can be used, it is valid, otherwise it is invalid.

If the adjacent block n is valid (step S4054: YES), the reference index refIdxLXN is set to the reference index refIdxLXn of the adjacent block n (step S4056). Also, the motion vector mvLXN is set as the motion vector mvLXn of the adjacent block n (step S4056), and the process of deriving the adjacent motion information of the block ends.

On the other hand, if the adjacent block n is invalid (step S4054: NO), the encoding information is acquired with the adjacent block n=B0 (step S4052), and it is determined whether the adjacent block n is valid or invalid (step S4054). . After that, similar processing is performed, and loops are performed in the order of B1 and A1. The process of deriving the adjacent motion information loops until the adjacent blocks become valid, and if all the adjacent blocks A0, B0, B1 and A1 are invalid, the process of deriving the adjacent motion information of the blocks ends.

Again, refer to FIG. After deriving the adjacent motion information (step S4004), temporal motion vectors are derived (step S4006).

Processing for deriving temporal motion vectors will be described with reference to FIG. First, the temporal motion vector tempMv=(0, 0) is initialized (step S4062).

Next, it is determined whether the adjacent motion information is valid or invalid (step S4064). If the flag availableFlagN=1 indicating whether or not the adjacent block can be used, it is valid, otherwise it is invalid. If the adjacent motion information is invalid (step S4064: NO), the process of deriving the temporal motion vector ends.

On the other hand, if the adjacent motion information is valid (step S4064: YES), it is determined whether or not the flag predFlagL1N indicating whether L1 prediction is used in the adjacent block N is 1 (step S4066). If predFlagL1N=0 (step S4066: NO), proceed to the next process (step S4078). If predFlagL1N=1 (step S4066: YES), it is determined whether the POCs of all pictures registered in all reference lists are less than or equal to the POC of the current picture to be processed (step S4068). If this determination is true (step S4068: YES), the process proceeds to the next process (step S4070).

If the slice type slice_type is B slice and the flag collocated_from_l0_flag is 0 (step S4070: YES, step S4072: YES), whether ColPic and reference picture RefPicList1[refIdxL1N] (picture with reference index refIdxL1N in reference list L1) are the same (step S4074). If this determination is true (step S4074: YES), the temporal motion vector tempMv=mvL1N (step S4076). If this determination is false (step S4074: NO), the process proceeds to the next process (step S4078). If the slice type slice_type is not B slice and the flag collocated_from_l0_flag is not 0 (step S4070: NO or step S4072: NO), the process proceeds to the next process (step S4078).

Then, it is determined whether or not the flag predFlagL0N indicating whether or not L0 prediction is used in adjacent block N is 1 (step S4078). If predFlagL0N=1 (step S4078: YES), it is determined whether ColPic and reference picture RefPicList0[refIdxL0N] (picture with reference index refIdxL0N in reference list L0) are the same (step S4080). If this determination is true (step S4080: YES), the temporal motion vector tempMv=mvL0N (step S4082). If this determination is false (step S4080: NO), the process of deriving the temporal motion vector ends.

Again, refer to FIG. Next, ColPic is derived (step S4016). Since this process is the same as S4201 in the temporal motion vector predictor candidate derivation unit 322, description thereof is omitted.

Then, coded blocks colCb at different times are set (step S4017). This sets, as colCb, a coded block positioned at the lower right of the center of the same position as the target coded block in the picture ColPic at a different time. This coding block corresponds to the coding block T1 in FIG.

Next, the position obtained by adding the temporal motion vector tempMv to the coding block colCb is set as a new colCb (step S4018). Assume that the upper left position of the coding block colCb is (xColCb, yColCb), and the temporal motion vector tempMv is (tempMv[0], tempMv[1]) with 1/16 pixel precision. and,
xColCb = Clip3( xCtb, xCtb + CtbSizeY + 3, xColCb + ( tempMv[0] >> 4 ) )
yColCb = Clip3( yCtb, yCtb + CtbSizeY - 1, yColCb + ( tempMv[1] >> 4 ) )
Calculate Here, the upper left position of the treeblock is (xCtb, yCtb), and the size of the treeblock is CtbSizeY. The coded block on ColPic containing position ((xColCb >> 3) << 3, (yColCb >> 3) << 3) becomes the new colCb. As shown in the above formula, the position after addition of tempMv is corrected within a range of about the size of the treeblock so as not to deviate greatly compared to before addition of tempMv. If this position is outside the screen, it is corrected inside the screen.

Then, it is determined whether or not the prediction mode PredMode of this coding block colCb is inter prediction (MODE_INTER) (step S4020). If the prediction mode of colCb is not inter prediction (step S4020: NO), set the flag availableFlagSbCol=0 indicating the presence of the sub-block temporal merge candidate (step S4003), and terminate the processing of the sub-block temporal merge candidate derivation unit. .

On the other hand, if the prediction mode of colCb is inter prediction (step S4020: YES), inter prediction information is derived for each reference list (steps S4022 and S4023). Here, for colCb, a central motion vector ctrMvLX for each reference list and a flag ctrPredFlagLX indicating whether or not LX prediction is used are derived. LX indicates a reference list, and in the derivation of reference list 0, LX becomes L0, and in the derivation of reference list 1, LX becomes L1. Derivation of inter prediction information will be described with reference to FIG.

If the coded block colCb at a different time cannot be used (step S4112: NO), or if the prediction mode PredMode is intra prediction (MODE_INTRA) (step S4114: NO), both flag availableFlagLXCol and flag predFlagLXCol are set to 0 (step S4116). , the motion vector mvCol is set to (0, 0) (step S4118), and the inter prediction information derivation process ends.

If the coding block colCb is available (step S4112: YES) and the prediction mode PredMode is not intra prediction (MODE_INTRA) (step S4114: YES), mvCol, refIdxCol and availableFlagCol are calculated in the following procedure.

If the flag PredFlagLX[xPCol][yPCol] indicating whether or not the LX prediction of the coding block colCb is used is 1 (step S4120: YES), the motion vector mvCol is the LX motion vector of the coding block colCb. MvLX[xPCol][yPCol] is set to the same value (step S4122), the reference index refIdxCol is set to the same value as the reference index RefIdxLX[xPCol][yPCol] of LX (step S4124), and the list listCol is set to LX. (step S4126). where xPCol, yPCol are indices indicating the position of the upper left pixel of the coded block colCb within the picture ColPic at different times.

On the other hand, when the flag PredFlagLX[xPCol][yPCol] indicating whether or not the LX prediction of the coding block colCb is used is 0 (step S4120: NO), the following processing is performed. First, it is determined whether or not the POCs of all pictures registered in all reference lists are less than or equal to the POC of the current picture to be processed (step S4128). In addition, it is determined whether or not the flag PredFlagLY[xPCol][yPCol] indicating whether or not the LY prediction of colCb is used is 1 (step S4128). Here, LY prediction is defined as a reference list different from LX prediction. That is, when LX=L0, LY=L1, and when LX=L1, LY=L0.

If this determination is true (step S4128: YES), the motion vector mvCol is set to the same value as MvLY[xPCol][yPCol], which is the LY motion vector of the encoding block colCb (step S4130), and the reference index refIdxCol is set to The same value as the reference index RefIdxLY[xPCol][yPCol] of LY is set (step S4132), and the list listCol is set to LX (step S4134).

On the other hand, if this determination is false (step S4128: NO), the flag availableFlagLXCol and the flag predFlagLXCol are both set to 0 (step S4116), the motion vector mvCol is set to (0, 0) (step S4118), and inter prediction information derivation processing is performed. exit.

When the inter prediction information is obtained from the coding block colCb, both the flag availableFlagLXCol and the flag predFlagLXCol are set to 1 (step S4136).

Subsequently, the motion vector mvCol is scaled to obtain a motion vector mvLXCol (step S4138). Since this process is the same as S4245 in the temporal motion vector predictor candidate derivation unit 322, description thereof is omitted.

Again, refer to FIG. After the inter prediction information is derived for each reference list, the calculated motion vector mvLXCol is set as the central motion vector ctrMvLX, and the calculated flag predFlagLXCol is set as the flag ctrPredFlagLX (steps S4022 and S4023).

Then, it is determined whether the central motion vector is valid or invalid (step S4024). If ctrPredFlagL0=0 and ctrPredFlagL1=0, it is determined to be invalid, otherwise it is determined to be invalid. If the center motion vector is invalid (step S4024: NO), the flag availableFlagSbCol=0 indicating the existence of the sub-block temporal merge candidate is set (step S4003), and the processing of the sub-block temporal merge candidate derivation unit is terminated.

On the other hand, if the central motion vector is valid (step S4024: YES), the flag availableFlagSbCol=1 indicating the existence of the sub-block temporal merge candidate is set (step S4025), and the sub-block motion information is derived (step S4026). This processing will be described with reference to FIG.

First, the number of sub-blocks numSbX in the width direction and the number of sub-blocks in the height direction numSbY are calculated from the width cbWidth and the height cBheight of the coding block colCb (step S4152). Also, refIdxLXSbCol=0 (step S4152). After this process, the process is repeated for each prediction sub-block colSb. This repetition is performed while changing the index ySbIdx in the height direction from 0 to numSbY and the index xSbIdx in the width direction from 0 to numSbX.

If the upper left position of the coding block colCb is (xCb, yCb), the upper left position (xSb, ySb) of the prediction sub-block colSb is
xSb = xCb + xSbIdx * sbWidth
ySb = yCb + ySbIdx * sbHeight
is calculated as Next, the position obtained by adding the temporal motion vector tempMv to the predicted sub-block colSb is set as a new colSb (step S4154). If the upper left position of the predicted sub-block colSb is (xColSb, yColSb) and the temporal motion vector tempMv is (tempMv[0], tempMv[1]) with 1/16 pixel precision, the upper left position of the new colSb is
xColSb = Clip3( xCtb, xCtb + CtbSizeY + 3, xSb + ( tempMv[0] >> 4 ) )
yColSb = Clip3( yCtb, yCtb + CtbSizeY - 1, ySb + ( tempMv[1] >> 4 ) )
becomes. Here, the upper left position of the treeblock is (xCtb, yCtb), and the size of the treeblock is CtbSizeY. As shown in the above formula, the position after addition of tempMv is corrected within a range of about the size of the treeblock so as not to deviate greatly compared to before addition of tempMv. If this position is outside the screen, it is corrected inside the screen.

Then, inter prediction information is derived for each reference list (steps S4156 and S4158). Here, for the prediction sub-block colSb, a motion vector mvLXSbCol for each reference list and a flag availableFlagLXSbCol indicating whether the prediction sub-block is valid are derived for each sub-block. LX indicates a reference list, and in the derivation of reference list 0, LX becomes L0, and in the derivation of reference list 1, LX becomes L1. The derivation of inter prediction information is the same as S4022 and S4023 in FIG. 47, so the description is omitted.

After deriving the inter prediction information (steps S4156, S4158), it is determined whether or not the prediction sub-block colSb is valid (step S4160). If availableFlagL0SbCol=0 and availableFlagL1SbCol=0, colSb is invalid, otherwise it is valid. If colSb is invalid (step S4160: NO), the motion vector mvLXSbCol is set as the center motion vector ctrMvLX (step S4162). Further, the flag predFlagLXSbCol indicating whether or not LX prediction is used is set to the flag ctrPredFlagLX in the central motion vector (step S4162). This completes the derivation of sub-block motion information.

Again, refer to FIG. Then, the L0 motion vector mvL0SbCol and the L1 motion vector mvL1SbCol are added as candidates to the subblock merge candidate list subblockMergeCandList in the subblock merge mode derivation unit 304 (step S4028). However, this addition is only when the flag availableSbCol=1, which indicates the presence of sub-block time merge candidates. Thus, the processing of the temporal merge candidate deriving unit 342 ends.

The above description of the sub-block temporal merging candidate deriving unit 381 is for encoding, but the same applies to decoding. In other words, the operation of the sub-block temporal merge candidate deriving unit 481 in the sub-block merging mode deriving unit 404 in FIG. 22 will be described in the same way, substituting decoding for encoding in the above description.

<Motion Compensation Prediction Processing>
The motion compensation prediction unit 306 acquires the position and size of the block currently being subjected to prediction processing in encoding. Also, the motion compensation prediction unit 306 acquires inter prediction information from the inter prediction mode determination unit 305 . A reference index and a motion vector are derived from the acquired inter prediction information, and the reference picture specified by the reference index in the decoded image memory is moved from the same position as the image signal of the prediction block by the amount of the motion vector. Generate a predicted signal after acquiring the signal.

When the inter prediction mode in inter prediction is prediction from a single reference picture such as L0 prediction or L1 prediction, the prediction signal obtained from one reference picture is the motion compensation prediction signal, and the inter prediction mode is BI When the prediction mode is prediction from two reference pictures, such as prediction, the weighted average of the prediction signals obtained from the two reference pictures is used as the motion-compensated prediction signal, and the motion-compensated prediction signal is used to determine the prediction method. Department. Although the ratio of weighted averaging of bi-prediction is set to 1:1 here, weighted averaging may be performed using other ratios. For example, the closer the picture interval between the picture to be predicted and the reference picture is, the higher the weighting ratio may be. Alternatively, the weighting ratio may be calculated using a correspondence table of combinations of picture intervals and weighting ratios.

The motion compensation prediction unit 406 has the same function as the motion compensation prediction unit 306 on the encoding side. The motion compensation prediction unit 406 passes the inter prediction information from the normal prediction motion vector mode derivation unit 401, the normal merge mode derivation unit 402, the sub-block prediction motion vector mode derivation unit 403, and the sub-block merge mode derivation unit 404 to switch 408. to get through.

The motion compensation prediction unit 406 supplies the obtained motion compensation prediction signal to the decoded image signal superimposing unit 207 .

<About inter-prediction mode>
The process of performing prediction from a single reference picture is defined as uniprediction, and in the case of uniprediction, either one of the two reference pictures registered in the reference lists L0 and L1, called L0 prediction or L1 prediction, is used. make predictions. L0 prediction and L1 prediction may be forward prediction (prediction referring to a forward reference image) or backward prediction (prediction referring to a backward reference image). 57 and 58 are diagrams for explaining motion compensation prediction in L0 prediction (uni-prediction).

FIG. 57 shows a case where the inter prediction mode is L0 prediction and the L0 reference picture (RefL0Pic) is before the processing target picture (CurPic). FIG. 58 shows a case in which the L0 prediction is performed and the L0 reference picture is at a time later than the picture to be processed. Similarly, uni-prediction can be performed by replacing the L0 prediction reference picture in FIGS. 57 and 58 with the L1 prediction reference picture (RefL1Pic).

A process of performing prediction from two reference pictures is defined as bi-prediction, and bi-prediction is expressed as bi-prediction using both L0 prediction and L1 prediction. 59 to 61 are diagrams for explaining motion compensation prediction in bi-prediction. FIG. 59 shows a bi-prediction case where the reference picture for L0 prediction is before the picture to be processed and the reference picture for L1 prediction is after the picture to be processed. FIG. 60 shows a bi-prediction case in which the reference picture for L0 prediction and the reference picture for L1 prediction are before the picture to be processed. FIG. 61 shows a bi-prediction case in which the reference picture for L0 prediction and the reference picture for L1 prediction are located after the picture to be processed.

In this way, the relationship between the prediction type and time of L0/L1 is limited to forward prediction (prediction referring to a forward reference image) for L0 and backward prediction (prediction referring to a backward reference image) for L1. It is possible to use it without In the case of bi-prediction, L0 prediction and L1 prediction may be performed using the same reference picture. It should be noted that the determination of whether motion compensation prediction is performed by uni-prediction or bi-prediction is made based on information (for example, a flag) indicating whether to use L0 prediction and whether to use L1 prediction, for example. be.

<About reference index>
In the embodiments of the present invention, it is possible to select an optimum reference picture from among a plurality of reference pictures in motion compensation prediction in order to improve the accuracy of motion compensation prediction. Therefore, the reference picture used in motion compensated prediction is used as a reference index, and the reference index is coded in the coded stream together with the coded vector.

<Motion Compensation Processing Based on Normal Predicted Motion Vector Mode>
Motion compensation prediction section 306, as also shown in inter prediction section 102 on the encoding side in FIG. obtains this inter prediction information from the inter prediction mode determination unit 305, derives the inter prediction mode, reference index, and motion vector of the block currently being processed, and generates a motion compensated prediction signal. The generated motion-compensated prediction signal is supplied to the prediction method determination unit 105 .

Similarly, motion compensation prediction section 406, as shown in inter prediction section 203 on the decoding side in FIG. Inter prediction information is obtained by the predicted motion vector mode derivation unit 401, and the inter prediction mode, reference index, and motion vector of the block currently being processed are derived, and a motion compensated prediction signal is generated. The generated motion-compensated prediction signal is supplied to the decoded image signal superimposing unit 207 .

<Motion Compensation Processing Based on Normal Merge Mode>
As shown in the inter prediction unit 102 on the encoding side in FIG. This inter-prediction information is acquired from the inter-prediction mode determination unit 305, the inter-prediction mode, reference index, and motion vector of the block currently being processed are derived, and a motion compensated prediction signal is generated. The generated motion-compensated prediction signal is supplied to the prediction method determination unit 105 .

Similarly, as shown in the inter prediction unit 203 on the decoding side in FIG. Inter prediction information is obtained by the derivation unit 402, the inter prediction mode, reference index, and motion vector of the block currently being processed are derived, and a motion compensated prediction signal is generated. The generated motion-compensated prediction signal is supplied to the decoded image signal superimposing unit 207 .

<Motion compensation processing based on sub-block prediction motion vector mode>
When the inter prediction mode determination unit 305 selects the inter prediction information by the sub-block prediction motion vector mode derivation unit 303, the motion compensation prediction unit 306 selects the inter prediction information by the inter prediction unit 102 on the encoding side in FIG. , this inter prediction information is obtained from the inter prediction mode determination unit 305, the inter prediction mode, reference index, and motion vector of the block currently being processed are derived, and a motion compensated prediction signal is generated. The generated motion-compensated prediction signal is supplied to the prediction method determination unit 105 .

Similarly, motion compensation prediction section 406, as shown in inter prediction section 203 on the decoding side in FIG. The sub-block prediction motion vector mode derivation unit 403 acquires inter prediction information, derives the inter prediction mode, reference index, and motion vector of the block currently being processed, and generates a motion compensation prediction signal. The generated motion-compensated prediction signal is supplied to the decoded image signal superimposing unit 207 .

<Motion Compensation Processing Based on Sub-Block Merge Mode>
As shown in the inter prediction unit 102 on the encoding side in FIG. 16 , the motion compensation prediction unit 306 performs , this inter prediction information is obtained from the inter prediction mode determination unit 305, the inter prediction mode, reference index, and motion vector of the block currently being processed are derived, and a motion compensated prediction signal is generated. The generated motion-compensated prediction signal is supplied to the prediction method determination unit 105 .

Similarly, as shown in inter prediction section 203 on the decoding side in FIG. Inter prediction information is acquired by the merge mode derivation unit 404, the inter prediction mode, reference index, and motion vector of the block currently being processed are derived, and a motion compensated prediction signal is generated. The generated motion-compensated prediction signal is supplied to the decoded image signal superimposing unit 207 .

<Motion Compensation Processing Based on Affine Mode>
Motion compensation using an affine model can be used in this embodiment. Motion compensation by the affine model uses 2 to 4 corners of an encoding block as control points, derives motion vectors of sub-blocks from the motion vectors of the control points, and performs motion compensation on a sub-block basis.

In this embodiment, in order to reduce the memory band in the affine mode, the division unit into sub-blocks is switched according to the condition of whether the motion information of the control point is bi-prediction or uni-prediction. In the case of uni-prediction, the sub-block size is set to 4×4 pixel units, and in the case of bi-prediction, the sub-block size is set to 8×4 pixel units. By restricting the small sub-block size in the case of bi-prediction prediction, it is possible to suppress the memory band when obtaining a reference image for motion compensation from memory. On the other hand, in uni-prediction, it is possible to improve coding efficiency by allowing smaller sub-block divisions to improve the accuracy of predicted images in uni-prediction.

When the motion information of the above control points is uni-predictive, the sub-block size is 8x4 pixel units, which is longer in the horizontal direction. effect can be obtained.

Processing for deriving motion vectors of sub-blocks from control points will now be described with reference to the flowchart of FIG.

It is determined whether or not the upper left control point CP0 is bi-predictive (step S1401). If CP0 is bi-predictive, set the sub-block horizontal size subBlockWidth and the sub-block vertical size as subBlockWidth=8 and subBlockHeight=4, respectively. Furthermore, the index offset values xSbIdxOffset and ySbIdxOffset for designating each subblock are set to xSbIdxOffset=2 and ySbIdxOffset=1, respectively (step S1402). If CP0 is not bi-predictive, that is, if it is uni-predictive, set the sub-block horizontal size subBlockWidth and the sub-block vertical size as subBlockWidth=4 and subBlockHeight=4, respectively. Furthermore, index offset values xSbIdxOffset and ySbIdxOffset for designating each sub-block are set to xSbIdxOffset=1 and ySbIdxOffset=1, respectively (step S1402). The values of subBlockWidth and xSbIdxOffset are different between steps S1402 and S1403.

The number of sub-blocks in the horizontal direction, numSubBlockWidh, and the number of sub-blocks in the vertical direction, numSubBlockHeight, are derived respectively. Here, if the horizontal size and vertical size of the encoding block to be processed are cbWidth and cbHeight, respectively, then numSubBlockWidth and numSubBlockHeight are
numSubBlockWidth = cbWidth / subBlockWidth,
numSubBlockHeight = cbHeight / subBlockHeight
(step S1403)
Motion vector variation amounts dHorX, dHorY, dVerX, dVerY of sub-blocks are calculated (step S1404). Here, dHorX represents the sub-block horizontal change amount of the horizontal component of the motion vector, and dHorY represents the sub-block vertical change amount of the horizontal component of the motion vector. Similarly, dVerX represents the sub-block horizontal variation of the vertical component of the motion vector, and dVerY represents the sub-block vertical variation of the horizontal component of the motion vector. Let dHorX, dHorY, dVerX, dVerY be
dHorX = (cp1MvX - cp0MvX) / (cbWidth/4)
dHorY = (cp1MvY - cp0MvY) / (cbWidth/4)
dVerX = (cp2MvX - cp0MvX) / (cbHeight/4)
dVerY = (cp2MvY - cp0MvY) / (cbHeight/4)
Derived by However, the motion vector of CP0 is (cp0MvX, cp0MvY), the motion vector of CP1 is (cp1MvX, cp1MvY), and the motion vector of CP2 is (cp2MvX, cp2MvY).

Each sub-block of the coding block is scanned in a predetermined order (steps S1405 and S1406) to derive the motion vector of each sub-block.

The reference positions xPosCb and yPoxCb for deriving the motion vectors of the sub-blocks are respectively
xPosCb = x + 0.5,
yPosCb = y + 0.5
Derived by x, y indicate the relative positional relationship when the upper left of the sub-block to be processed is (0, 0).

Add 0.5 to x, y to derive the motion vector reference position xPosCb, yPosCb of the sub-block, and derive the motion vector mvSubX(x, y), mvSubY(x, y) of the sub-block (x, y) (step S1407).
mvSub(x,y) and mvSuv(x,y) are
mvSub(x,y) = cp0MvX + dHorX*xPosCb + dHorY*yPosCb,
mvSuv(x,y) = cp0MvY + dVerX*xPosCb + dVerY*yPosCb
Derived by Derive mvSub(x, y), mvSuv(x, y) for all sub-blocks and terminate the procedure.

FIG. 66(a) shows the coordinate values (x, y) and the motion vector reference positions xPosCb, yPosCb of the sub-blocks when the 16×16 size encoding block is divided into 4×4 sub-blocks, and FIG. Coordinate values (x, y) obtained by dividing a 16×16 coded block into 8×4 sub-blocks and motion vector reference positions xPosCb and yPosCb of the sub-blocks are shown. In this embodiment, when divided into 4x4 subblocks, the motion vector reference positions xPosCb and yPosCb of the subblocks are centered in the 4x4 subblocks, as shown in FIG. 66(a). On the other hand, when divided into 8x4 sub-blocks, as shown in FIG. 66(b), the motion vector reference positions xPosCb and yPosCb of the sub-blocks are not at the center of the 8x4 sub-blocks, and xPosCb when divided into 4x4 sub-blocks. , takes the same position as the yPosCb position.

Although 8x4 sub-blocks and 4x4 sub-blocks have been described in this embodiment, the sub-block size is not limited to this. The reference position of the motion information of the M×N sub-blocks coincides with the reference position of one N×N sub-block, ie square area, when the M×N block is divided into a plurality of N×N sub-blocks, ie square areas.

By adopting the configuration of this embodiment, the following effects are obtained.

Since the reference position for deriving the motion vector of the subblock does not change even when the division size of the subblock changes, there is an effect that the processing for deriving the motion vector of the subblock can be simplified.

By simplifying the processing for deriving the motion vector of the sub-block, it becomes possible to switch the division size appropriately according to the image characteristics, and the coding efficiency can be improved.

The following flags are reflected in the following flags based on the inter-prediction conditions determined by the inter-prediction mode determination unit 305 in the encoding process, and encoded in the encoded stream. In the decoding process, whether or not to perform motion compensation by the affine model is specified based on the following flags in the encoded stream.

sps_affine_enabled_flag indicates whether motion compensation by an affine model can be used in inter prediction. If sps_affine_enabled_flag is 0, motion compensation by the affine model is suppressed on a per-sequence basis. Also, inter_affine_flag and cu_affine_type_flag are not transmitted in the CU (Coding Unit) syntax of the coded video sequence. If sps_affine_enabled_flag is 1, affine model motion compensation can be used in the encoded video sequence.

sps_affine_type_flag indicates whether motion compensation by 6-parameter affine mode can be used in inter prediction. The 6-parameter affine model is a mode in which motion vectors of sub-blocks are derived from six parameters of horizontal and vertical components of motion vectors of three control points, respectively, and motion compensation is performed on a sub-block basis. A motion vector is derived in units of subblocks, but a common reference index is derived in units of coding blocks.

If sps_affine_type_flag is 0, motion compensation is not performed by the 6-parameter affine model. Also, cu_affine_type_flag is not transmitted in the CU syntax of the encoded video sequence. If sps_affine_type_flag is 1, motion compensation with a 6-parameter affine model can be used in the encoded video sequence.

If sps_affine_type_flag is not present, it shall be 0.

When decoding a P or B slice, if inter_affine_flag is 1 in the CU currently being processed, the affine model is used to generate the motion-compensated prediction signal of the CU currently being processed. Motion compensation is used.

If inter_affine_flag is 0, the affine model is not used for the CU currently being processed.

If inter_affine_flag is not present, it shall be 0.

When decoding a P or B slice, if cu_affine_type_flag is 1 in the CU currently being processed, 6-parameter affine Motion compensation by model is used.

If cu_affine_type_flag is 0, then motion compensation with a 4-parameter affine model is used to generate a motion compensated prediction signal for the CU currently being processed. The 4-parameter affine model is a mode in which a sub-block motion vector is derived from four parameters, ie, horizontal and vertical components of motion vectors of two control points, and motion compensation is performed on a sub-block basis.

<Merge Differential Motion Vector (MMVD)>
A differential motion vector can be added to the motion vectors of the top two merge candidates (merge candidates with merge indices of 0 and 1 in the merge candidate list). This differential motion vector is called a merge differential motion vector.

When the merge difference motion vector is added in the encoding-side merge candidate selection unit 347 , the motion vector to which the merge difference motion vector is added is supplied to the motion compensation prediction unit 306 via the inter prediction mode determination unit 305 . Also, the bitstream encoding unit 108 encodes information about the merge difference motion vector. The information about the merge difference motion vector is the index mmvd_distance_idx indicating the distance to be added to the motion vector and the index mmvd_direction_idx indicating the direction to add the motion vector. These indices are defined as in the tables shown in FIGS. 63(a) and 63(b). Then, if the x and y components of the merge differential motion vector offset MmvdOffset are represented by MmvdOffset[0] and MmvdOffset[1] respectively,
MmvdOffset[0] = ( MmvdDistance << 2 ) * MmvdSign[0]
MmvdOffset[1] = ( MmvdDistance << 2 ) * MmvdSign[1]
becomes. The merge differential motion vector is derived from the merge differential motion vector offset MmvdOffset in the above equation. The details of deriving the merge difference motion vector are described below in the case of the decoding side.

On the decoding side, if a merge differential motion vector exists, information on the merge differential motion vector is separated from the bitstream supplied to the bitstream decoding unit 201, and a merge differential motion vector offset MmvdOffset is derived. Also, the merge candidate selection unit 447 derives a merge differential motion vector from the decoded merge differential motion vector offset. After adding this merged differential motion vector to the motion vector, the motion vector is supplied to the motion compensated prediction unit 406 .

Derivation of the merge differential motion vector mMvdLX in the merge candidate selection unit 447 will be described with reference to the flowchart of FIG. 64(a). First, it is determined whether or not the inter prediction mode of the coding block is bi-prediction (PRED_BI) (S4402). If it is not bi-prediction (S4402: No), it is determined whether it is L0 prediction (PRED_L0) (S4404). In the case of L0 prediction (S4404: Yes),
mMvdL0 = MmvdOffset
mMvdL1 = 0
(S4406), the process of deriving the merge difference motion vector ends. For L1 prediction (S4404: No),
mMvdL0 = 0
mMvdL1 = MmvdOffset
(S4408), the process of deriving the merge difference motion vector ends.

On the other hand, in the case of bi-prediction (S4402: Yes), the difference between the POC of the picture to be processed currPic and the reference picture is calculated for each reference list and set to currPocDiffL0 and currPocDiffL1 (S4410). Here, the POC difference between picA and picB, DiffPicOrderCnt(picA, picB), is
DiffPicOrderCnt( picA, picB ) = [POC of picA] - [POC of picB]
indicates Reference picture RefPicList0[refIdxL0] is a picture indicated by reference index refIdxL0 of reference list L0. Similarly, reference picture RefPicList1[refIdxL1] is a picture indicated by reference index refIdxL1 in reference list L1.

Next, it is determined whether -currPocDiffL0*currPocDiffL1>=0 (step S4412). If this determination is true (step S4412: Yes),
mMvdL0 = MmvdOffset
mMvdL1 = -MmvdOffset
As (step S4414), the process of deriving the merge difference motion vector ends. On the other hand, if this determination is false (step S4412: No),
mMvdL0 = MmvdOffset
mMvdL1 = MmvdOffset
(step S4416). Next, it is determined whether or not the absolute value of the POC difference from the reference list L0 is greater than or equal to the absolute value of the POC difference from the reference list L1 (step S4418). If this determination is true (step S4418: Yes), set X=0, Y=1 (step S4420), and scale the merged differential motion vector mMvdL1 of L1 (step S4424). Here, mMvdLY indicates mMvdL0 when Y=0 and mMvdL1 when Y=1. On the other hand, if this determination is false (step S4418: No), set X=1, Y=0 (step S4422), and scale the merged differential motion vector mMvdL0 of L0 (step S4424). The scaling of the merge differential motion vector mMvdLY is as shown in FIG. 64(b).
td = Clip3( -128, 127, currPocDiffLX )
tb = Clip3( -128, 127, currPocDiffLY )
tx = ( 16384 + Abs( td ) >> 1 ) / td
distScaleFactor = Clip3( -4096, 4095, ( tb * tx + 32 ) >> 6 )
mMvdLY = Clip3( -32768, 32767, Sign( distScaleFactor * mMvdLY )
* ( (Abs( distScaleFactor * mMvdLY ) + 127 ) >> 8 ) )
derived as Here, currPocDiffLX indicates currPocDiffL0 when X=0 and currPocDiffL1 when X=1. Similarly, currPocDiffLY indicates currPocDiffL0 if Y=0 and currPocDiffL1 if Y=1. Clip3(x,y,z) is a function that limits the minimum value to x and the maximum value to y for the value z. Sign(x) is a function that returns the sign of the value x, and Abs(x) is a function that returns the absolute value of the value x. With the above, the process of deriving the merge difference motion vector ends.

The merge difference motion vector may be added to the motion vectors of the top two sub-block merge candidates. In this case, the index mmvd_distance_idx indicating the distance to be added to the motion vector is defined as in the table shown in FIG. 63(c). Since the operation of the sub-block merge candidate selection unit 386 is the same as that of the merge candidate selection unit 347, description thereof will be omitted. Also, the operation of the sub-block merge candidate selection unit 486 is the same as that of the merge candidate selection unit 447, so the description is omitted.

As described above, MmvdDistance is defined as in the tables shown in FIGS. 63(a) and 63(c). Since these tables are defined with quarter-pixel precision, the generated merge difference motion vectors may contain fractional-pixel precision. However, by encoding/decoding the flag indicating that the pixel precision in these tables is 1 in units of slices, the generated merge difference motion vector can be changed so that it does not include fractional pixel precision. .

<Adaptive motion vector resolution (AMVR)>
The resolution of the differential motion vector can be adaptively changed for each encoding block. This resolution is called adaptive motion vector resolution.

A case of using the adaptive motion vector resolution for the normal motion vector predictor mode will be described. In this case, the spatial motion vector predictor candidate deriving units 321 and 421, the temporal motion vector predictor candidate deriving units 322 and 422, and the historical motion vector predictor candidate deriving units 323 and 423 derive the candidate motion vectors according to the resolution. is rounded. The resolution can be selected from 1/4, 1, and 4 pixel precision, and if the resolution is not changed, it will be 1/4 pixel precision. The rounding process is performed according to the motion vector resolution in the encoding block to be processed. That is, the derived candidate motion vector mvX is
rightShift = leftShift = MvShift + 2
offset = 1 << ( rightShift - 1 )
mvX = ( mvX >= 0 ? ( mvX + offset ) >> rightShift :
- ( ( - mvX + offset ) >> rightShift ) ) << leftShift
and rounded. Here, MvShift=0 when the resolution of the motion vector in the encoding block to be processed is 1/4 pixel precision. Similarly, MvShift=2 if the motion vector resolution is 1-pixel precision, and MvShift=4 if the motion vector resolution is 4-pixel precision. The above formula handles each of the x,y components of mvX.

Adaptive motion vector resolution can also be used for sub-block predictive motion vector mode. In this case, only the resolution differs from the normal motion vector predictor mode described above. That is, in the affine inherited motion vector predictor candidate deriving units 361 and 461, the affine constructed motion vector predictor candidate deriving units 362 and 462, and the affine identical motion vector predictor candidate deriving units 363 and 463, the derived candidate motion vectors are rounded according to . The resolution can be selected from 1/16, 1/4, and 1 pixel precision, and if the resolution is not changed, it will be 1/16 pixel precision. The rounding process is performed according to the motion vector resolution in the encoding block to be processed. That is, the derived candidate motion vector mvX is rounded according to the above formula. Here, MvShift=0 when the resolution of the motion vector in the encoding block to be processed is 1/4 pixel precision. Similarly, if the motion vector resolution is 1-pixel precision, MvShift=2. The above equations process each of the x,y components of mvX.

In all the embodiments described above, the encoded bitstream output by the image encoding device has a specific data format so that it can be decoded according to the encoding method used in the embodiments. are doing. The encoded bitstream may be provided by being recorded in a computer-readable recording medium such as an HDD, SSD, flash memory, or optical disk, or may be provided from a server through a wired or wireless network. Therefore, an image decoding device corresponding to this image encoding device can decode the encoded bitstream of this specific data format regardless of the providing means.

When a wired or wireless network is used to exchange an encoded bitstream between an image encoding device and an image decoding device, the encoded bitstream is converted into a data format suitable for the transmission mode of the communication channel. may be transmitted. In this case, a transmission device that converts an encoded bitstream output from an image encoding device into encoded data in a data format suitable for the transmission mode of a communication channel and transmits the encoded data to a network, and a transmission device that receives encoded data from the network. and a receiving device for restoring an encoded bitstream to supply to an image decoding device. The transmission device includes a memory that buffers the encoded bitstream output from the image encoding device, a packet processing unit that packetizes the encoded bitstream, and a transmission unit that transmits the packetized encoded data via the network. including. The receiving device includes a receiving unit that receives packetized encoded data via a network, a memory that buffers the received encoded data, and packetizes the encoded data to generate an encoded bitstream, and a packet processing unit provided to the image decoding device.

When a wired or wireless network is used to exchange coded bitstreams between an image coding device and an image decoding device, in addition to the transmitting device and the receiving device, the coded data transmitted by the transmitting device is also used. A relay device may be provided for receiving and supplying to the receiving device. The relay device includes a receiver that receives packetized encoded data transmitted by the transmitter, a memory that buffers the received encoded data, and a transmitter that transmits the packetized encoded data and the network. include. Further, the relay device includes a reception packet processing unit that performs packet processing on packetized encoded data to generate an encoded bitstream, a recording medium that stores the encoded bitstream, and a packetization of the encoded bitstream. A transmission packet processing unit may be included.

Further, the display device may be configured by adding a display unit for displaying an image decoded by the image decoding device. In this case, the display unit reads the decoded image signal generated by the decoded image signal superimposition unit 207 and stored in the decoded image memory 208 and displays it on the screen.

Further, an imaging device may be configured by adding an imaging unit to the configuration and inputting a captured image to an image encoding device. In that case, the imaging unit inputs the image signal of the imaged image to the block dividing unit 101 .

The above encoding and decoding processes may of course be implemented as transmission, storage, and reception devices using hardware, and are stored in ROM (read only memory), flash memory, or the like. It may be implemented by firmware or software such as a computer. The firmware program or software program may be recorded on a computer-readable recording medium and provided, or may be provided from a server through a wired or wireless network, or data broadcasting of terrestrial or satellite digital broadcasting. can be provided as

The present invention has been described above based on the embodiments. It should be understood by those skilled in the art that the embodiments are examples, and that various modifications can be made to combinations of each component and each treatment process, and such modifications are also within the scope of the present invention. .

100 image coding device 101 block division unit 102 inter prediction unit 103 intra prediction unit 104 decoded image memory 105 prediction method determination unit 106 residual signal generation unit 107 orthogonal transformation/quantization unit 108 bit string code 110 decoded image signal superimposition unit 111 encoded information storage memory 200 image decoding device 201 bit string decoding unit 202 block division unit 203 inter prediction unit 204 intra prediction unit , 205 encoded information storage memory, 206 inverse quantization/inverse orthogonal transform unit, 207 decoded image signal superimposition unit, 208 decoded image memory.

Claims (1)

A video encoding device that divides an encoding block into subblocks, derives a motion vector for each subblock, and performs affine encoding,
Obtaining a plurality of control points having motion vectors, determining reference positions of the sub-blocks, deriving motion vectors of the sub-blocks based on the control points and the reference positions, and performing motion compensation on a sub-block basis. a motion compensated prediction unit,
The motion compensation prediction unit divides the sub-blocks into non-square sub-blocks that are rectangular areas that are not square,
A video encoding device, wherein a reference position for deriving a motion vector of the sub-block is set as a reference position of one of the square regions obtained by dividing the non-square sub-block into a plurality of square regions. .

Images